pSeven 6.28

Released November 2021. Includes pSeven Core 6.28.

Updates and Changes

  • Predictive modeling:
    • core-158 Improved time limit implementation in SmartSelection (the Training time limit… hint) to provide a better kind of time-quality trade-off. Lowering the time limit now hints the SmartSelection algorithm to use models that are less accurate but train faster, which allows it to try more approximation techniques within a limited time. This potentially yields better results than training internal models with maximum quality but limiting the set of used techniques, as in previous versions.
  • Design space exploration block:
    • core-141 Improved automatic selection of levels for continuous variables in the Full Factorial technique. Levels are now selected with the aim to generate a full factorial design that has size as close as possible to the specified Exploration budget value. The technique is now able to select a different number of levels for each continuous variable. For example: in previous versions, the Full Factorial technique always selected the same number of levels for each continuous variable, so for a design with 5 continuous variables and requested point count of 242 (\(3^5-1\)) it generated 32 points (\(2^5\)), because it could select only 2 levels for each of the 5 variables. New version in this case generates 216 points, variables in the resulting design have varying number of levels (\(4 \times 3 \times 3 \times 3 \times 2 = 216\)).
    • core-141 Improved selection of levels for continuous variables in the Orthogonal Array technique. The technique now selects a varying number of levels for continuous variables in order to generate an orthogonal or balanced array with size as close as possible to the Exploration budget value. Previously it often generated a full factorial DoE with a minimum number of levels.
  • ApproxBuilder block:
    • p7-220 When configuring the block in SmartSelection mode, you can now use the new input port x_categorical to specify categorical variables. Previously you had to add the Advanced options… hint to specify categorical variables.
  • STAR-CCM+ block:
    • p7-240 The block is promoted to release blocks and is no longer tagged as beta.
  • CSVParser block:
    • p7-192 The block no longer provides the Remove empty fields option, which often caused unwanted effects. Empty fields are now treated as missing values, which are represented by special None values if the output is RealMatrix, and empty strings "" if the output is StringMatrix.

Compatibility Issues

  • This release disables a controversial CSVParser feature — the option to remove empty fields found in the file from the output data (Remove empty fields). In many cases, enabling this option led to an error because data could not be output as a rectangular matrix, since removing empty fields created rows or columns with different length. To avoid unexpected behavior, CSVParser now always keeps empty fields, representing them as missing values in the output matrix (None in RealMatrix, empty strings "" in StringMatrix). This change may create issues in rare cases where removing empty fields did not interfere with matrix output. If you have used CSVParser to load data with missing values, in particular from files with missing rows or columns, it is recommended to review and adjust block settings after upgrading to pSeven 6.28.


  • p7-220 Added the x_categorical port description to the ApproxBuilder block documentation (see section Training Data).


  • Predictive modeling:
    • core-158 Fixed an issue with using the GBRT technique in SmartSelection training where it stopped prematurely after training a candidate GBRT model, if you did not use a test sample.
    • core-185 Fixed an issue where internal validation information was missing from a model trained in the partial linear output dependency mode, if the training sample contained duplicates.
    • core-158 Fixed an issue where parallel threads sometimes were not used in training despite parallelization is enabled.
  • Design space exploration block:
    • core-141 Fixed various issues with categorical variables in the Adaptive design for GP technique.
    • core-141 Fixed various issues with discrete variables in the D-optimal design and I-optimal design techniques.
    • core-141 Fixed an issue where the Parametric study technique incorrectly selected levels for continuous variables, using one of the variable bounds instead of a central point placed between bounds.
  • General:
    • p7-222 Fixed an issue where some blocks set to idle run mode could cause the workflow run to fail — for example, integration type blocks performed unnecessary search for program libraries, which may be unavailable.
    • p7-128 Fixed an issue where pSeven displayed project description with a noticeable delay when opening a project with many open workflows (saved session). Project description is now loaded before re-opening the workflows from the previous session.
    • p7-51 Fixed an issue where pSeven did not open some links from its project description. Links to file types associated with applications other than pSeven now open in the respective applications as expected.

pSeven 6.27

Released October 2021. Includes pSeven Core 6.27.

New Features

  • p7-198 The new SolidEdge block (beta) enables integrating Siemens Solid Edge parts and assemblies into pSeven workflows to apply optimization and other design space exploration techniques to them. Note that using the SolidEdge block requires an updated license file.

Updates and Changes

  • Predictive modeling:
    • core-122 Improved support of categorical inputs and outputs in models exported to executable or C program source. Exported models can now work with string values of categorical variables and also support loading samples from CSV files.
    • core-122 Models exported to executable or C program source will no longer print gradient values by default. To output gradients, add the -d option to the model’s command line.


  • p7-223 Added the SolidEdge block documentation.


  • Design space exploration block:
    • p7-221 Fixed an issue in using the Adaptive design technique with categorical variables, due to which the block produced incorrect results in tasks where no feasible designs can be found for some value combinations of categorical variables — for example, it could output an empty matrix to Feasible designs even though feasible designs are found for other value combinations.
  • Predictive modeling:
    • core-174 Fixed an issue where training a model with partial linear output dependency mode caused an error if point weights were specified.
    • core-146 Fixed an issue in SmartSelection where updating a model sometimes failed because pSeven could not train a new model with specified options and hints.
    • core-160 Fixed several issues with training a model with categorical outputs, which could cause errors in training.
    • core-122 Fixed several issues related to using string values with models that have categorical inputs or outputs.
    • core-122 Fixed an issue where Unicode names of model variables could cause malformed log messages and other errors in logging.

pSeven 6.26

Released September 2021. Includes pSeven Core 6.26.

New Features

  • p7-47 The Design space exploration block is updated with a new algorithm to handle linear responses (constraints, in particular) in the Adaptive design, Surrogate-based optimization, and Gradient-based optimization techniques. If you set function type of a response to linear, this is now interpreted as a requirement that changes the block behavior as follows:

    • The block always aims to fit each response configured as linear with a linear model. Linear response modeling is now performed prior to starting the main optimization or adaptive design algorithm. It requires a certain relatively low number of linear response evaluations, which depends on the number of design variables.
    • If the response cannot be fitted with a linear model, it means that the linearity requirement is not met, and in this case the block raises an error, which is then processed according to the Error handling behavior option.
    • If response modeling succeeds, the block continues to use the obtained response model internally. This guarantees that the Design space exploration block will not request any more response evaluations from the blackbox block that calculates linear response values.

    This update also resolves the issue where the Surrogate-based optimization or Adaptive design technique was unable to create an internal model of some linear response but continued to evaluate it, treating that response as a computationally cheap function, which allows an unlimited number of evaluations.

Updates and Changes

  • Run:
    • p7-112 Simplified user interface for manually stopping blocks: it now uses a single stop button that automatically applies the appropriate function (stop or terminate) depending upon the block type; extra dialogs were removed.
  • ApproxBuilder block:
    • p7-184 The block can now generate an HTML report containing model information and validation data — see section Model Report in the block documentation for details.


  • p7-102 Updated section Workflow with descriptions of workflow execution controls and logging functions (see subsections Run Management and Run Logs).
  • p7-112 Updated descriptions of block stop functions as per user interface changes — see section Manual Stop.
  • p7-138 Updated section Initial Samples in the Design space exploration block documentation with more details about using different kinds of initial sample ports.
  • p7-47 Updated descriptions of discrete variables, linear and quadratic responses in the Design space exploration block documentation.
  • p7-204 Updated the ApproxBuilder block documentation to explain how to get an HTML report containing model information and validation data — see Model Report.
  • p7-114 Updated section Model Validator with more details on test data settings.


  • Predictive modeling:
    • core-173 Fixed an issue with internal validation where specifying size of validation subsets (GTApprox/IVSubsetSize) could lead to an error if the training sample contains ambiguous points (different response values for same inputs).
  • Design space exploration block:
    • p7-20 Fixed the Unsolvable problem behavior option in optimization techniques not working as intended.
    • p7-213 Fixed an issue where you could not switch a variable to constant using the <variable_name>.Constant port, if the variable was not marked as constant in block configuration.
    • core-106 Fixed stability issues in surrogate-based optimization algorithms, which appeared when applying global search methods to problems where local search methods prove to provide a more accurate solution.
    • core-137 Fixed an issue in the Surrogate-based optimization technique where it ignored the Computational scalability option in multi-objective problems, or when the Local search option was enabled.
    • p7-138 Fixed several issues in validation of initial samples, which could lead to errors if the sample data received to component-wise initial sample ports is inconsistent — for example, there are missing input components, or component-wise samples have different sizes.
    • p7-138 Fixed misleading error messages about incorrect usage of component-wise initial sample ports.
    • core-147 Fixed an issue with the Adaptive design for GP technique where it could raise an error if the initial sample consists of a single point or contains a constant input column.
  • General:
    • p7-202 Fixed excessive logging in the pSeven batch executable (p7batch).

pSeven 6.25

Released August 2021. Includes pSeven Core 6.25.

New Features

  • p7-188 The Adaptive design technique in the Design space exploration block now supports stepped variables.
  • p7-124 When you train models in Analyze, progress bars in the Models pane now show percent estimates of training progress. Progress bar tooltips provide additional information — for example, the number of completed training phases.

Updates and Changes

  • Design space exploration block:
    • core-119 Improved quality of sample-based Adaptive design (with disabled blackboxes) where responses have adaptive type.
    • p7-187 Added logging of initial sample pre-processing operations in Adaptive design.
  • Predictive modeling:
    • core-116 Improved SmartSelection algorithms to avoid performance degradation when training large models.
    • core-120 Improved tuning of GBRT and MoA technique parameters in SmartSelection.
    • p7-185 When you export a model to Octave using an ApproxBuilder or an Approximation model block, the block validates model filename and function name and shows a warning on invalid names.
  • General:
    • p7-76 pSeven uninstaller now shows a confirmation dialog if you start uninstallation while pSeven is running. Note that if you accept uninstallation in this case, it requires a reboot to complete.



  • Analyze:
    • p7-167 Fixed an issue with the Dependency tab in Sample viewer where guidelines on the scatter plot disappeared when comparing negative values.
    • p7-190 Fixed an error in Model validator, which sometimes appeared when you change models.
  • Predictive modeling:
    • core-125 core-116 Fixed issues in SmartSelection where it sometimes did not interrupt training when stopped, or the Training time limit… hint was not working as intended.
    • core-140 core-94 Fixed several issues with model update, which appeared in models with categorical inputs or outputs.
  • Design space exploration block:
    • core-149 Fixed an issue with the Adaptive design technique where it ignored the initial sample in tasks with categorical or discrete variables.
    • core-132 Fixed an issue where the block could sometimes incorrectly qualify an optimization problem as unsolvable if it received some NaN values of responses.
  • Program block:
    • p7-178 Fixed inconsistent naming of the Success codes option in block configuration dialog.
  • CSVParser block:
    • p7-186 Fixed an issue where the block could not read a CSV file with empty fields.

pSeven 6.24

Released July 2021. Includes pSeven Core 6.24.

Updates and Changes

  • Predictive modeling:
    • core-52 Model export to C and Octave now generates less code.
  • Design space exploration block:
    • core-127 Improved Surrogate-based optimization (SBO) technique behavior when the Computational scalability option is specified so that sizes of design batches output for evaluation are better adjusted to the option value. Note that it does not guarantee that design batch size is a multiple of Computational scalability at every SBO iteration.
  • PythonScript block:
    • p7-123 Environment variable settings from the Environment tab now apply only if you configure the block to use an external Python interpreter (see External Python).
    • p7-123 Added a hint on the Environment tab explaining the use of environment variable settings from this tab.
  • Program block:
    • p7-123 Added a hint on the Environment tab explaining the use of environment variable settings from this tab.
  • CSVParser block:
    • p7-163 If the CSV file contains string data, the block now automatically reads and outputs it as a StringMatrix, so you are no longer required to set the Data type option manually. The Data type option is removed.


  • p7-101 Updated section Configuration in the Design space exploration block documentation to explain the Evaluation budget and Study target settings in Adaptive design.
  • p7-107 Updated section Environment Settings in the Program block documentation.
  • p7-107 Updated section Environment Settings in the PythonScript block documentation.
  • p7-107 Removed section Environment from Block Configuration since environment settings are available only in the Program and PythonScript blocks.
  • p7-116 Updated section Known Issues in the Excel block documentation with a recommendation to use Exit Sub instead of the End statement in Excel macros.


  • Analyze:
    • p7-158 Fixed an issue in the Sample viewer where error values were not sorted together with cell values when sorting the data table by columns.
  • Predictive modeling:
    • core-107 Fixed an issue with multithreading on processors with a high number of cores (10 or more) which caused training performance degradation compared to pSeven 6.16.
    • core-96 Fixed an issue where you could not update a GBRT model if the initial model was trained with a sample that contains duplicates.
  • Design space exploration block:
    • p7-172 Fixed an issue with the Adaptive design technique where an infeasible problem caused duplicate evaluations, if the Unsolvable problem behavior option was set to “Output all designs”.
  • PythonScript block:
    • p7-123 Fixed an issue where environment variable settings from the Environment tab did not apply when testing the script. Note that environment settings apply only if you configure the block to use an external Python interpreter (see External Python).
    • p7-123 Fixed an issue where variable substitutions on the the Environment tab (using ${VARIABLE} syntax) did not work.

pSeven 6.23

Released June 2021. Includes pSeven Core 6.23.

New Features

  • p7-130 Initial support for classification: pSeven can now train models with categorical outputs, which work as classifiers. A categorical output takes values from a predefined set — that is, the set of class labels (note that labels should be numeric). Values of a categorical output predicted by a model also come from this set — that is, the model predicts class membership for an input sample. To train a classification model, select categorical outputs in the Build model dialog when training a model in Analyze.

Updates and Changes

  • Analyze:
    • p7-139 You can now export multiple models to a single Excel document: just select the models in the Models pane and use the Export to file… command from the pane’s b_context menu.
    • p7-155 Sample viewer can now sort rows by index, so you can quickly reset row order to default when needed.
  • Predictive modeling:
    • core-89 Improved SmartSelection algorithms to avoid loss of model quality in cases where the training sample is a Cartesian product of several factor sets (as in Tensor Approximation).
    • core-114 Improved SmartSelection algorithms to avoid performance degradation when training models with a high number of input variables.
    • core-114 When set manually, the GTApprox/Accelerator option now affects RSM parameter estimation in SmartSelection.
  • Design space exploration block:
    • p7-86 New settings for the Adaptive design technique: you can now set the number of feasible designs to generate, or the maximum number of designs to evaluate, or both. Due to this, the Number of designs setting is replaced with two new parameters: Exploration budget, which sets the design evaluation limit, and Study target, which sets the target number of feasible designs. The Study target setting is currently supported for the Adaptive design technique only. Note that Design space exploration blocks in workflows created in previous versions of pSeven are upgraded automatically but some of them may get equal Exploration budget and Study target values, which produces a warning — see the block’s Issues pane for details.
  • ANSA block:
    • p7-143 The block no longer stops immediately when it cannot launch ANSA due to not enough license credits available. Instead, the block will keep retrying to capture license and start ANSA, until it exceeds the timeout set by the ANSA license timeout option.
  • CSVParser block:
    • p7-126 To avoid confusion, the Merge field delimiters option is renamed to Remove empty fields — this name correctly describes the option behavior. Also, this option is now disabled by default, so empty CSV fields are not removed. If the sample file contains empty fields, the block outputs the sample as a RealMatrix where corresponding elements are None (or empty strings, if you set output type to StringMatrix).



  • Predictive modeling:
    • core-103 Fixed an issue where using a training sample with named inputs and outputs (columns) caused an error if the names contain local language characters.
  • Design space exploration block:
    • core-104 Fixed the Adaptive design technique behavior in cases where an adaptive response evaluates to NaN.
    • core-91 Fixed misleading error messages when using an initial sample in tasks with stepped variables.
  • STAR-CCM+ block:
    • p7-161 Fixed an issue where a STAR-CCM+ block that works in some cycle (for example, optimization) started failing on every cycle iteration after a one-time error in the STAR-CCM+ application.
  • CSVParser, CSVGenerator blocks:
    • p7-126 Fixed various issues with empty CSV fields and samples that contain missing values. The blocks now properly support empty values.
  • General:
    • p7-144 Fixed pSeven uninstaller sometimes not removing the pSeven installation directory completely.

pSeven 6.22

Released May 2021. Includes pSeven Core 6.22.

New Features

  • p7-93 Added the idle run feature for blocks, which you can use to skip execution of a block — for example, when testing a workflow. Idle blocks do not process input data they receive, and their output ports send only values assigned to them — or nothing, if there is no value assigned. To set a block to idle, select it in the workflow and use the Idle run command from the b_context menu in the workflow edit toolbar.

Updates and Changes

  • Run:
    • p7-137 When you set a run name, you can now use the %s substitution for the name of the currently selected parameter preset — in addition to the previously available %d and %t substitutions for the current date and time.
  • Design space exploration block:
    • p7-131 To avoid confusion, columns corresponding to responses with disabled evaluation (the No blackbox hint) are now included in the matrices output to All designs and other result ports — so matrix columns always match variables and responses defined in the block configuration.
  • Text block:
    • p7-90 Improved quick selection behavior to generate fewer Set frame start… and Set frame end… operations.

Compatibility Issues

  • This release changes matrix shapes in Design space exploration block results in the case when its configuration includes responses with the No blackbox hint: columns corresponding to such responses are no longer removed from results. You may need to update workflows and reports in projects where Design space exploration blocks solve tasks with disabled responses.


  • Run:
    • p7-141 Fixed an issue where a workflow appeared as running indefinitely, and you could not stop it or quit pSeven normally.
  • Analyze:
    • p7-108 Fixed an issue where Model explorer showed plots in such a way that plot curves did not pass through the origin point specified in its configuration.
  • Predictive modeling:
    • core-67 Fixed a bug in the GBRT training technique, which caused an error when training a GBRT model with many categorical variables with large sets of levels.
  • PythonScript block:
    • p7-118 Fixed an error that occurred when you use an external Python interpreter to run a script that works with structured NumPy arrays with custom data types.
  • General:
    • p7-92 Fixed an issue where pSeven did not respect the system settings for the decimal delimiter when you copy or paste floating point values — for example, in Sample viewer, workflow inputs or outputs.
    • p7-137 Fixed an issue where pSeven could generate an invalid name for an automatically created port — for example, when you configure an integration type block and select nodes from its parameter tree to map them to ports.

pSeven 6.21

Released April 2021. Includes pSeven Core 6.21.

Updates and Changes

  • Analyze:
    • p7-70 Added the command to quickly switch a report between results from different workflow runs — see Select Source Run.
  • Predictive modeling:
    • core-26 Increased performance and improved stability of the PLA training technique.
    • core-26 Added extrapolation support to models trained using the PLA technique.
  • Design space exploration block:
    • core-63 Improved the Adaptive design for GP technique to minimize the RRMS error of Gaussian processes (GP) models trained using a DoE generated by this technique.
    • p7-46 Cleaned up the block run logs printed to Info. Internal messages from the optimizer are now visible only in Debug. Also, the block now reports the number of points in intermediate results to the run log.
  • Creo block:
    • p7-111 Updated the block for compatibility with Creo and above. Note that versions and are not supported due to issues in Creo.


  • p7-98 Added the ANSA block documentation.
  • p7-104 Fixed several outdated screenshots and explanations in Tutorials.


  • Analyze:
    • p7-15 Fixed an issue with dataset configuration in various viewers where you could not re-add a data source after deleting it from the dataset.
  • Predictive modeling:
    • p7-127 Fixed performance issues in ApproxBuilder and other modeling blocks, caused by a bug that appeared in pSeven 6.17 and effectively disabled training in parallel mode.
    • core-75 Fixed an issue with model validation in SmartSelection, which negatively affected its performance when using holdout test data to estimate model quality.
  • Uncertainty quantification block:
    • p7-84 Fixed a bug in the Reliability analysis (SORM) technique, which caused an error when the block configuration includes variables with different order of magnitude.
    • p7-117 Fixed an issue with the Simulation technique, which caused an error when the block could not fit the simulated sample to some of the distributions.
  • Text block:
    • p7-83 Fixed an issue where it was possible to configure a read operation so it converts a value to a data type that is not compatible with the variable (port) type, causing errors.
  • STAR-CCM+ block:
    • 25 Fixed an issue where the block was unable to find the STAR-CCM+ executable for versions 14 or higher.

pSeven 6.20

Released March 2021. Includes pSeven Core 6.20.

New Features

  • p7-29 The new ANSA block (beta) enables integrating ANSA models with optimization tasks into pSeven workflows. The block reads task configuration and builds a parameter tree, which includes task’s design variables, responses, histories, and model’s measurements and mass properties. You can then select parameters from the tree to map them to block’s inputs and outputs. Note that using the ANSA block requires an updated license file.
  • 19237 Parallel execution support for the STAR-CCM+ block. STAR-CCM+ now supports running multiple simulations in parallel, when it is nested into a parallel Composite block. Several STAR-CCM+ blocks in the same workflow can also run in parallel according to their configuration — see the Allow parallel block execution option for details. This type of parallelization is possible only when the block launches STAR-CCM+ locally and is not compatible with the block’s remote execution features.

Updates and Changes

  • Analyze:
    • p7-67 A few convenience features in Model validator:
      • You can now specify a test sample only for one or a few model outputs you validate. The tool no longer requires configuring a test sample for every output.
      • If you select some data series in the Report database pane before adding a Model validator to a report, pSeven automatically adds the selected data series to test samples.
  • Uncertainty quantification block:
    • p7-72 In the simulation mode, the block now properly notifies that response bounds are ignored by graying them out and adding a tooltip.
  • Creo block:
    • p7-37 Updated the block for compatibility with Creo 6.0.
  • DoE block (deprecated since 6.12):
    • 35 Removed the legacy DoE block from the block library. All supported DoE methods are available in the Design space exploration block.


  • p7-103 Fixed a few outdated screenshots in Tutorials.

Compatibility Issues

  • This release removes the legacy DoE block from the block library. If you open a workflow containing a DoE block in pSeven 6.20, it automatically deletes the block from the workflow. To avoid issues, it is recommended to replace DoE blocks with Design space exploration before upgrading the project to 6.20.


  • Design space exploration block:
    • p7-95 Fixed a minor issue with SmartSelection due to which it sometimes selected the Gradient-based optimization technique when this was not intended.
  • ApproxBuilder block:
    • p7-97 Fixed an issue with block upgrade, due to which pSeven 6.19 could not open the block’s configuration dialog when editing a workflow created in previous versions of pSeven.
  • General:
    • p7-79 Fixed incorrect behavior of the source selector in the pSeven console after you clear the console log.

pSeven 6.19

Released February 2021. Includes pSeven Core 6.19.

New Features

  • p7-63 core-34 Added a more convenient way to set up leave-n-out cross validation in ApproxBuilder and the Model builder in Analyze in the manual configuration mode. Use the GTApprox/IVSubsetSize option to set n — for example, to perform leave-one-out cross-validation. Use the updated GTApprox/IVTrainingCount option to limit the number of model retraining sessions. See option descriptions for full details.

Updates and Changes

  • Design space exploration block:
    • core-25 Improved uniformity and variability of DoE generated by the Adaptive design technique in the uniform distribution mode (no adaptive responses) and in certain tasks with discrete variables.
  • Predictive modeling:
    • core-9 The GBRT technique now handles NaN values of variables in the training sample in a specific way. If you set the GTApprox/InputNanMode option, it keeps points where some (but not all) variables are NaN and actually uses them in training.



  • Design space exploration block:
    • p7-34 Fixed an issue with initial sample processing in Surrogate-based optimization, where the technique could violate linear constraints due to errors occurring when training internal response models based on the initial sample.
    • p7-20 p7-82 Fixed an issue where the block sent True to the done output even if its error handling behavior was configured otherwise. Note that since 6.16.1, the error raised in case of an infeasible problem is by default suppressed by the Unsolvable problem behavior option, and the block outputs evaluated designs and an empty solution instead of raising an error. If you have earlier used the Error handling behavior option to treat infeasible problems, it is recommended to edit your workflow to process the block’s results according to the new default behavior. Alternatively, you can set Unsolvable problem behavior to Technique error to reproduce old behavior.
    • core-16 Fixed some issues in validation of variable bounds, which rarely led to incorrectly qualifying the problem as infeasible.
  • Predictive modeling:
    • core-17 Fixed an issue where a TA or iTA model training failed if GTApprox/TAModelReductionRatio is 1 and exact fit is required. Reduction ratio 1 means no reduction, so the exact fit requirement is valid in this case.
    • core-54 Fixed a crash when smoothing an SGP model trained on noisy data with given output noise variance.
    • core-19 Fixed the inability to export some data fusion (.gtdf) models to C#. Note that C# export requires an up to date license valid for pSeven 6.16 and above.
  • PythonScript block:
    • p7-40 Fixed incorrect block behavior when testing a script with List or Dict variables, which contain missing components (RealScalar data with the special None value).
  • Analyze:
    • p7-80 Fixed issues with Model builder where you could not stop training or view logs during training.

pSeven 6.18

Released January 2021. Includes pSeven Core 6.18.

New Features

  • Extended Python support in PythonScript and user blocks enables using Python 3 and various Python environments with pSeven.
    • p7-13 The PythonScript block can now run external Python instead of the default Python version built into pSeven. See section External Python for details.
    • p7-26 It is also possible to use custom Python version to run your user block code — see section Block Implementation for details.
  • Version upgrade for user blocks — see Version Compatibility and Update in User Blocks for details. Note that user blocks from pSeven 6.17 have to be upgraded to get the extended Python support.

Updates and Changes

  • Design space exploration block:
    • rd-19491 When generating a DoE with the Orthogonal array design technique, you can now set a strict requirement for the array type using the new Array type requirement option.
    • core-13 Stability improvements in local search algorithms in Surrogate-based optimization.
  • Uncertainty quantification block:
    • 19255 Added the second order reliability analysis method (SORM).
    • 19255 Reports output by the block now contain distribution parameters and a hidden timestamp.
  • Predictive modeling:
    • core-4 core-33 core-37 core-18293 The Tensor Approximation (TA) technique now supports piecewise linear approximation (PLA) as a technique that you can use for TA factors. In particular, this enables TA to provide n-linear approximations on a grid.
    • core-18423 Improved compatibility of exported model code, added more validity checks to avoid compilation issues.
    • core-18293 Updated training sample size requirements for the HDA, GP, HDAGP, SGP, and MoA techniques. See the pSeven Core manual for details.
  • Edit:
    • rd-19469 Added the user tag to user blocks in the Block library pane. Note that user blocks from pSeven 6.17 have to be upgraded to get the tag. Also added the command to show or hide user blocks to the Block library pane’s b_context menu.
    • rd-19258 Further improvements in workflow auto-layout functions, continuing from pSeven 6.17.
  • Analyze:
    • rd-19200 Plots in Model explorer now show markers for every point evaluated when exploring the model, even if the point is outside the training sample domain (input bounds).
  • General:
    • p7-18880 rd-19135 Various minor usability improvements and updates.



  • Workspace:
    • rd-17208 Fixed an error when copying files with names containing some special characters.
  • Analyze:
    • p7-22 Fixed column sort issues in Sample viewer with samples that contain NaN values or columns of different length.
    • rd-19454 Fixed Model explorer not reporting model training errors correctly sometimes.
    • rd-19214 Fixed an issue that caused incorrect behavior of various viewers if you were making changes in a viewer’s configuration while it loads data.
    • rd-19474 Fixed plot legend settings in the Export image dialog in the Column chart and Histogram viewers.
  • Design space exploration block:
    • rd-19477 Fixed an issue with initial sample processing in tasks with discrete and stepped variables, where the block could filter out some of the actually valid points from the sample, if the sample contained data with limited precision — for example, loaded from a CSV or text file.
  • Predictive modeling:
    • core-18796 Fixed an extrapolation issue in some MoA models, which returned 0 everywhere outside of the training sample domain.
    • core-10 Fixed an issue where evaluating a MoA model exported as an Octave script (.m) raised an error if the input point is outside of the training sample domain.
  • General:
    • p7-44 Fixed an issue with startup checks for temporary files directories due to which pSeven sometimes did not start properly.
    • rd-19475 Fixed an issue where pSeven Agent’s help did not open in Linux.
    • rd-19475 Fixed an issue in pSeven Agent’s compatibility with Ubuntu 20.04.
    • rd-19468 rd-19484 Various minor usability and GUI fixes.

pSeven 6.17

Release date: 20 November 2020.

This release includes pSeven Core (MACROS) 6.17.


With the release of pSeven 6.17, the support for 32-bit platforms is finally canceled due to lack of demand. pSeven releases do not include 32-bit installers by default since version 6.2, and there was no interest in 32-bit builds. Future versions of pSeven will support only 64-bit Windows and Linux editions.

New Features

  • 19103 19208 19461 19451 User blocks. pSeven provides a full-featured block template, which you can use to extend the pSeven’s block library with new integrations. A new custom block generated from the template works out-of-the-box: it already contains a graphical interface, where the user can select parameters to create input and output ports — similarly to other integration type blocks, and also supports error handling, sandbox directories, and other common block features. The amount of required custom code is minimal: the block just needs 2 Python functions, which build the tree of parameters and process the input data, to wrap any software tool you need, add a graphical interface to an existing Python script, add support for a new model format, and so on. See section User Blocks for full details on available customizations and a development guide.
  • Design space exploration block:
    • 19136 Stepped variables support in optimization. A stepped variable sticks to the grid defined by the variable’s bounds and step size. Stepped variables are intended for the common case of a continuous quantity (geometry parameter, load, power), which can be changed only in certain increments due to various practical reasons. See section Types of Variables in the block’s documentation for details.
    • 19136 Variables with resolution in optimization studies. Specifying resolution for variables essentially sets precision of optimization techniques, adjusting their algorithms for a particular task. Compared to running optimization with maximum numerically possible resolution, setting proper resolution typically results in improved runtime (less response evaluations) at the cost of insignificant change in solution quality. Resolution is set using a new variable hint — see Variable Hints for details.
  • 18958 19107 More control over model training technique settings in SmartSelection. Previously, you could either run SmartSelection in the default mode, where it selects the best technique automatically, or choose a single specific technique to use, so SmartSelection tunes training settings for this technique only. In pSeven 6.17, you can use the new EnabledTechniques hint to allow two or more certain techniques; SmartSelection then selects the best among them, ignoring the techniques that you disabled. This feature is supported in the ApproxBuilder block and in Analyze. See section Training Modes for configuration details.
  • 18797 18807 18946 18954 19083 19131 19132 19157 19165 Many internal improvements in approximation techniques used by SmartSelection, and several stability and performance updates in SmartSelection algorithms. Effectiveness of these updates depends on the training data and optional SmartSelection settings. In most of the typical usage scenarios, they result in higher model quality an a noticeable increase in training performance.
  • More additions to predictive modeling tools:
    • 18568 18896 18939 19500 Added the support for importing data fusion models into reports. You can now use data fusion models (.gtdf, .df) with all predictive modeling tools available in Analyze similarly to approximation models (.gtapprox, .gta).
    • 18974 Export a dimension reduction model (.gtdr, .dr) as an Excel document with a linked DLL. This export feature is available through the DRCompressor and DRDecompressor blocks: load a model to the block, then use the export command from the Model file pane. Note that export to Excel is supported only in the Windows version of pSeven.
    • 17845 18957 When training models with the ApproxBuilder block, you can add specific input and output bounds. Input bounds define the model’s input domain (valid ranges of inputs); the block can determine them from the training data, or you can adjust the bounds manually. Output bounds add thresholds for model outputs. See the descriptions of the sample_lower_bounds and sample_upper_bounds input ports in section Training Data for details.
    • 16548 You can now hard stop model training in Analyze using the Terminate training command. See Model Training for details.
    • 18316 Updated the Model validator in Analyze:
      • The details table now contains 2 tabs: Errors with model accuracy statistics and Points with full details on validation points. You can also select points of interest on the scatter plot to filter the list on the Points tab.
      • The scatter plot in the Model validator now supports linked selection (see Linked Selection).
      • Tooltips on the scatter plot show both input and output values.
      • A few more interface improvements.

Updates and Changes

  • Workspace:
    • 18883 Several safety updates related to project database backup and restore. When you open a project where the database file is corrupt, and pSeven cannot fix it automatically, it now shows a database repair dialog. In the dialog, you can select to restore the database from a backup, clear it, or cancel opening the project to keep the database intact. When automatic repair succeeds, pSeven also keeps a copy of the corrupt database file, so you can undo the repair if needed. Database backups created during a workflow run are now additionally verified for integrity.
    • 18908 Added better project lock checks when opening a project. It should no longer be possible to unlock a project currently used by another pSeven instance, even if the project is located on a network drive or is opened by a pSeven user from another host. The project unlock dialog in this case contains details on the host and process which are using the project.
  • Edit:
    • 18296 pSeven now shows a warning if you link a block that works inside a loop to a block outside of the loop. This workflow pattern often leads to an error when running the workflow, because the block outside of the loop cannot accept multiple outputs generated by the loop.
    • 18869 To avoid confusion, it is no longer possible to export the workflow root block using the Export workflow command, since in this case export actually creates a copy of the current workflow.
    • 18881 Minor usability improvements in the workflow import dialog, which opens when you import an existing workflow from a file as a Composite block in another workflow.
    • 17494 18989 18587 Improved the workflow auto-layout functions.
  • Analyze:
    • 17015 The Sample viewer can now sort the table on the Data tab by any column in the increasing or decreasing value order, similarly to spreadsheet editors.
    • 18064 Reworked the image export dialogs in tools that support plot export. Added options to adjust image size, select file format, and others.
    • 18496 Added number format settings for axis tick labels on 2D plots.
    • 18949 When you use the Rename… command to rename multiple data series, it now finds both the common prefix and suffix in names.
    • 18792 The Correlations tab in the Sample viewer now allows copying correlation coefficients and p-values regardless of the selected analysis mode.
    • 18326 When you add data monitored from the Design space exploration block ports to a report, pSeven now automatically removes the “Blackbox” prefix from names of data series for convenience, so the names of data series are the same as the names of block’s variables and responses.
    • 18972 When you add a matrix that contains columns of different data types from a project database to a report, pSeven now converts columns to numeric data series when possible. This update is important when working with results of a Design space exploration block that defines some string (categorical) variables: in this case its result matrices contain both string and numeric columns, which were previously all converted to string data series, causing inconvenience.
    • 17477 Updated the general data processing mechanisms in order to increase performance when adding large amounts of data from a project database to a report.
    • 19210 18776 Clarified several error messages related to data import and export.
  • Design space exploration block:
    • 19494 Added more robust checks for the initial sample data passed to the Surrogate-based optimization and Adaptive design techniques. When analyzing the initial sample, these techniques now update it as needed to ensure that the sample properly covers the design space defined by variable bounds. In particular, this helps in avoiding unwanted localization in tasks where the initial sample is clustered in a relatively small area of the design space.
    • 18802 19247 19259 19238 19515 Updated the local search algorithms of the Surrogate-based optimization technique, improving their overall stability and performance. Also, you can now specify a starting point for the forced local search mode (the Local search option set to Forced): just set the initial guess in variable properties.
    • 19136 The Gradient-based optimization and Surrogate-based optimization techniques now require that ranges between bounds of variables are at least \(10^{-6}\) to avoid several stability issues.
    • 18911 Added a warning on using discrete variables with linear and quadratic responses for the Adaptive design, Gradient-based optimization, and Surrogate-based optimization techniques.
    • 19196 Noticeably increased performance of the Latin hypercube sampling technique when it is used to generate big samples (thousands of points).
    • 19166 Improved the support for categorical and discrete variables in the Latin hypercube sampling technique, resulting in better sample point distribution when generating a DoE with a combination of continuous and discrete or categorical variables.
    • 19147 Improved the Latin hypercube sampling technique results quality when it is used with an initial sample.
    • 18914 Significantly increased the Orthogonal array technique performance and stability for big samples and designs where variables have many levels (about 10 or more).
    • 19136 Updated the block’s graphical interface to accommodate the increasing number of variable and response properties. Secondary properties are now set using hints — see Editing Variables and Responses, Variable Hints, and Response Hints for details.
  • Uncertainty quantification block:
    • 19140 Reports generated by the block now contain more details on analysis variables, including the type and parameters of their distributions.
  • Predictive modeling:
    • 19083 19165 Updated the GP and HDAGP techniques to improve their stability, in particular, better avoid overtraining in certain cases.
    • 18598 You can now use the MoA technique to update a model trained using the GBRT technique.
    • 19010 18551 In the Model details dialog, you can now export model constraints to a text file. Also, the output constraints formula is now better optimized for brevity.
    • 18928 You can now copy model output sensitivity scores (Sobol indices) calculated by the Model explorer to the clipboard or export them to a text file, using commands from the b_blconf_context menu.
    • 17398 The point of view in the Model explorer is now set to one of the model’s training points by default.
    • 18434 To make retraining a model on its training sample (stored to the model) more convenient, the Extract training sample… command now uses exact names of model inputs and outputs as the names of the new data series it creates, so the Retrain… command recognizes them automatically.
    • 19053 When you export an approximation model as an Excel document with a linked DLL, names of model evaluation functions are now validated to avoid conflict with the names of built-in Excel functions. Previously, such name clashes could cause compilation errors.
    • 19006 The Build model, Retrain model, and Update model dialogs no longer contain the column that specifies the test data filter, as this feature is mostly unused. In case you need to filter a test sample, you can load it into a Sample viewer, apply filters and use the Create new data series… command to save filtered data.
    • 19165 Changed the formatting of model accuracy and internal validation information printed to the training log for better readability.
  • ApproxBuilder block:
    • 18568 18896 18939 19500 You can now use a data fusion model (.gtdf, .df) as an initial model in training. In this case, the block updates a data fusion model to an approximation model, improving it with the new training data.
    • 19132 Added the Auto setting for the GTApprox/GPType, GTApprox/RSMFeatureSelection, GTApprox/RSMType, and GTApprox/SPLTContinuity options. This value is now default. This is intended for better compatibility with SmartSelection; in the manual training mode, the Auto setting behavior is the same as the old default.
  • Approximation model block:
    • 18568 18896 18939 19500 Added the support for data fusion models (.gtdf, .df). You can load a data fusion model to the block and use it similarly to an approximation model (.gtapprox, .gta).
    • 18340 The lists of input and output bounds in the Model summary pane now use names of inputs and outputs which were saved to the model, instead of default numbered names. Model summary also lists categorical variables, if any. The same updates are done to the Model summary section in reports generated by the block.
    • 18282 Input and output descriptions saved to the model are now used as variable and response port descriptions by default.
  • Excel block:
    • 18491 Added the Startup attempts option to control the number or retries when connecting to Excel.
  • ANSYSWorkbench block:
    • 18891 Added the support for the Optimize Update Order feature in ANSYS Workbench. Also added the Optimize update order option, which you can use to enable or disable this feature.
  • STAR-CCM+ block:
    • 19101 Added the support for STAR-CCM+ version 14.
    • 18959 The block now supports all STAR-CCM+ licensing schemas, including the support for Lite and Power licenses. See License Settings for details.
    • 18959 Added the STAR-CCM+ License.License path option to specify the location of the STAR-CCM+ license file or connect to a non-default license server.
    • 18569 Increased the level of detail in block’s logs.
  • Program block:
    • 19194 When you open the block, its configuration dialog now takes less time to load.
    • 18919 Renamed the Script tab to Executable as it is used to configure a program command line as well as a shell script.
  • Text block:
    • 18020 Increased responsiveness of the block’s configuration dialog when working with large templates.
    • 18924 The block now shows warnings more clearly when you test operations.
    • 19091 The block can now handle sequences of Read value… operations that fill different components of a vector or matrix variable. It requires you to specify a test value for the variable to provide information on vector length or matrix shape to the block. See section Testing Operations for more details.
  • PythonScript block:
    • 19179 You can now use Java libraries in PythonScript thanks to the jpype Python module added to pSeven.
    • 17 Added statistics to built-in modules.
  • Composite block:
    • 19060 Added options that set up caching. You can add these options to workflow parameters, making block’s cache settings available in Run. See section Cache in the Composite block documentation for details.
  • CSVParser block:
    • 18775 Updated behavior of the Import rows and Merge field delimiters settings to make it consistent with similar settings in the data exchange tools in Analyze.
  • General:
    • 19074 On startup, pSeven now checks that the system directory for temporary files is configured properly. If the temporary files directory is not accessible, pSeven does not start and shows a message with error details, because such access errors are known to cause various issues when running workflows.
    • 17382 18411 18791 18861 18885 18889 18964 18985 18990 18993 19005 19038 19201 Various usability improvements and updates.



  • Workspace:
    • 19453 Fixed an issue where pSeven did not resolve a name conflict, if you copy a report to a project that already contains a report with the same name.
  • Edit:
    • 18786 Fixed various errors that occurred if you use a command hotkey (for example, Del to remove a block) while moving a block, link, and in other similar cases. pSeven will now correctly disable hotkeys during drag-and-drop actions in Edit.
    • 16149 Fixed an issue where you could delete a block from the workflow while opening this block’s configuration dialog.
    • 18832 Fixed an issue where pSeven remembered the block name once specified in the Add block dialog and then prohibited to use this name even though you closed the dialog without adding the block.
    • 18925 Fixed an issue where uplinking two ports with complex data types to the same input or output port resulted in a general error.
  • Run:
    • 19463 Fixed the bug that allowed changing parameter presets while a workflow is running.
    • 19458 Fixed an issue where pSeven could start a wrong workflow or print the workflow run log to another tab, if you quickly switch workflow tabs after clicking the b_run button.
    • 18593 Fixed several minor issues with parameter values displayed incorrectly after you reset them.
  • Analyze:
    • 19206 Fixed an issue where pSeven could needlessly reload all plots in a report after you rename some of its data series.
    • 19066 Fixed a bug due to which pSeven showed a change marker in the title of a report tab right after you open that report.
    • 18936 Fixed a bug due to which data export to Excel or CSV files could work incorrectly when export settings specify the same name for 2 or more columns.
    • 18930 Fixed an issue where you could specify an invalid starting cell when exporting data to Excel.
    • 18808 Fixed some issues with adding data from a project database to a report, if the data was monitored from a port of the List type.
    • 19146 Fixed an error in the Sample viewer when exporting sample statistics to a file.
    • 18912 Fixed the table on the Data tab in the Sample viewer resetting its horizontal scroll position after you apply a value filter to a column.
    • 19240 Fixed a rare issue where the Sample viewer’s Data tab showed invalid contents after you scroll the table.
    • 19497 Fixed the inputs selector on the Dependency tab in the Sample viewer showing incorrect names of inputs after you rename related data series in the report database.
    • 19063 Fixed an issue where sensitivity scores (Sobol indices) calculated by the Sample viewer were inconsistent with the scores calculated by the Model explorer.
    • 18988 Fixed an issue in 2D plot where selecting a plot area and double-clicking the plot activated zoom instead of removing the selection.
    • 18262 Fixed 3D plots not showing the coordinate grid.
  • Design space exploration block:
    • 19154 Fixed an issue with the Adaptive design technique where it could not handle a linear constraint as intended, if response data for the constraint is available from the initial sample, and new evaluations are disabled for the constraint (the response has no blackbox).
    • 18915 18911 19501 Fixed several cases of incorrect behavior of the Adaptive design, Gradient-based optimization, and Surrogate-based optimization techniques in tasks with discrete variables.
    • 18808 Fixed an issue where the block could not write results to the project database, if it used an initial data sample of the IntMatrix type.
    • 19129 Fixed an issue where using a technique that does not support discrete or categorical variables and runs exploration for each possible combination of levels of these variables consumed all available memory, if the number of possible combinations was high enough.
    • 18808 Fixed the bug resulting in the inability to collect data from the block if it works with an initial sample supplied as an integer matrix.
    • 19142 Fixed an issue with the Initial sample evaluation technique, which sometimes caused an error, if the block is configured to remove duplicates from the initial sample.
    • 19153 Fixed a few issues with the Adaptive design for GP technique, which negatively affected its performance, stability, and result quality.
    • 19227 Fixed an issue in the Latin hypercube sampling technique, which sometimes lead to generating a DoE with a non-uniform marginal distribution.
  • Uncertainty quantification block:
    • 19163 19243 Fixed several issues caused by specifying invalid distribution parameters for a variable.
    • 19470 Fixed the block sometimes not saving settings specified in the Ports and parameters dialog.
    • 19182 Fixed an issue with the sample-based multinomial distribution type, which caused an error when running the workflow.
  • Predictive modeling:
    • 18599 Fixed an issue in the Model explorer where the input and output selectors sometimes did not show their names properly.
    • 18920 Fixed unexpected slowdowns in the Model validator caused by changing plot settings or adding models.
    • 18475 19069 Fixed various issues in the Model explorer and Model validator caused by deleting one of the loaded models, or deleting a model and then adding another one under the previously used name.
    • 19244 Fixed an error that sometimes occurred when exporting an RSM model as an Excel document with a linked DLL.
    • 19041 18571 Fixed a few errors and warnings that appeared when exporting some approximation models.
    • 19111 19512 Fixed a few issues with model export to Octave.
    • 18857 Fixed the duplicated line breaks in model details exported to a text file.
    • 18782 Fixed a GUI error when simultaneously training multiple models, which appeared if one of the models failed to train, and you tried to stop training of another model.
    • 18820 Removed deprecated options from the Training options dialog in Analyze, which is used to configure model training manually.
    • 19203 Fixed an issue where model training configured in the manual mode could freeze, if the training sample contains many NaN output values, and the GTApprox/OutputNanMode option is set to "Predict".
    • 17879 Fixed an issue where model details contained incorrect information about constant outputs.
    • 18807 Fixed the support for output transformation (see GTApprox/OutputTransformation) in the PLA technique.
    • 18946 Fixed a bug that rarely lead to incorrect calculation of model accuracy and internal validation statistics, if the training sample contains many NaN output values.
    • 18807 Fixed an issue where the training log contained incorrect values of the GTApprox/OutputTransformation option.
    • 18807 Fixed the GTApprox/InputNanMode option compatibility with the SPLT approximation technique.
    • 19455 Fixed a bug in the Update model dialog that disabled switching between the GBRT and MoA techniques when updating a GBRT model, which is in fact compatible with both these techniques.
    • 18901 Fixed an issue with invalid ApproxBuilder block option values when opening workflows created in previous pSeven versions.
  • DFBuilder, DRBuilder blocks:
    • 19081 Fixed errors that occurred when running, if the block is configured to receive the model path to the model_path port.
  • Excel block:
    • 18825 Fixed the ability to specify an invalid cell range when adding or editing variables.
    • 18845 Fixed cell indices in block configuration being case-sensitive.
    • 18491 Fixed invalid block behavior when running an Excel macro that closes Excel.
    • 18069 Fixed a few issues where validation warnings in block configuration were interpreted as validation errors, so you could not start the workflow.
  • ANSYSWorkbench block:
    • 19516 Fixed an issue where the block could not build the parameter tree, if the project contains integer parameters or parameters with special types. The block will now skip parameters of unsupported types, issuing a warning to the console log.
  • FMI model block:
    • 18794 Fixed an issue with temporary files which resulted in an error printed to the console log when you run the workflow.
  • STAR-CCM+ block:
    • 19067 Fixed some issues with geometry assemblies imported to STAR-CCM+ from other CAD tools.
  • Program block:
    • 18987 Fixed an issue where the block could not output a large file to a port and continued to run indefinitely without any warnings.
    • 18852 Fixed an issue where you could specify an invalid exit code in the list of success exit codes, which caused an error when running the workflow.
    • 19011 When you use the remote port to specify SSH connection settings, the block no longer requires you to save the SSH username to configuration.
    • 18788 18850 Fixed a few interface issues.
  • Text block:
    • 19074 Fixed an error that occurred when the block is unable to write to the system directory for temporary files.
    • 18921 Fixed the Variables pane showing an incorrect test value of a StringVector variable when testing a sequence of Read value… operations that fill different components of the vector.
    • 19075 Fixed an issue with the operation test function where it could not determine the line number containing an error and reported a list index error instead of the actual error.
    • 18486 Fixed a bug in the Add variables dialog due to which it sometimes opened with the “Name” field already containing a name.
  • PythonScript block:
    • 18601 Fixed the Script output pane not printing exception details properly or reporting an incorrect line number, when an error occurs while testing the script.
    • 18977 Fixed an error when using the Test script command to test a script that imports a Python module built into pSeven.
  • Composite block:
    • 18593 Fixed some inconsistencies in the port type and value inheritance rules in the case when there are multiple nested block ports uplinked to the same input port of their parent Composite block.
  • CSVParser block:
    • 19085 Fixed an error occurring when you specify an invalid row range to the Import rows setting.
    • 19170 Fixed the syntax for specifying the names of columns to import not working as intended.
  • General:
    • 18929 Fixed the fmpy-32 compilation error message in the pSeven installer which appeared on some Linux systems when finishing installation.
    • 19000 Fixed an issue where pSeven could not exit correctly after it failed to close a project due to a project database error.
    • 19168 Fixed a bug which could cause a crash when exiting pSeven.
    • 18991 Fixed the warnings that appeared in the pSeven console when opening reports, which were created in previous pSeven versions.
    • 18535 18866 18872 19459 19090 18856 19119 18846 18559 18858 18868 Various minor usability and GUI fixes.

pSeven 6.16.4

Release date: 4 September 2020.

This release includes pSeven Core (MACROS) 6.16.4.

New Features

  • 19103 This release provides early access to one of the upcoming major updates — the support for developing your own blocks. See section User Blocks for details.



  • General:
    • 19236 Fixed an issue with workflows containing blocks that are not supported or not found in the block library of the current pSeven installation, and are therefore removed when opening the workflow. Previously pSeven automatically saved the workflow after removing such blocks, so you could not restore the workflow to its original state, unless you had a backup copy of the project. Now pSeven removes unsupported blocks but does not save the workflow, so you can close it without saving to cancel changes.

pSeven 6.16.3

Release date: 13 April 2020.

This release includes pSeven Core (MACROS) 6.16.3.

Updates and Changes

  • Design space exploration block:
    • 18983 Increased performance of the Surrogate-based optimization (SBO) technique, resulting in noticeably faster problem solving.
    • 19019 Increased optimization techniques performance in problems with a high number (about 500 or more) of variables or constraints.
    • 19047 19057 19079 Improved the Adaptive design technique performance, stability and quality in certain kinds of tasks, in particular when generating a uniform adaptive DoE with linear constraints.
    • 19118 The Local search activation delay option of the Surrogate-based optimization (SBO) technique was removed as it has been found that delayed localization does not provide any significant advantage over a pure local SBO (the default behavior in the forced local search mode), while introducing needless complexity in configuration. Now if the Local search option is set to \(Forced\), the local search mode activates immediately. Blocks in workflows created in pSeven versions 6.16 to 6.16.2, which have the Local search activation delay option set, are automatically updated to the new behavior: essentially, the delay is set back to \(0\) (default).
  • Predictive modeling:
    • 19036 Improved stability of several approximation techniques, including MoA and GBRT.
  • CSVParser block:
    • 19046 CSV files with empty lines are now handled correctly, and empty lines are ignored. Note that the first line must not be empty if the block is configured to parse it as a table header.


  • Analyze:
    • 19035 19049 Fixed an issue where some viewers in a report did not load due to a license error.
    • 19035 19049 Fixed error handling when opening reports. Now if some part of a report fails to load, pSeven skips it and continues loading other data instead of stopping and showing an empty report.
  • Design space exploration block:
    • 19089 Fixed a deadlock issue in the case when the Constant property of some variable is mapped to an input port and there is a link connected to this port.
    • 19118 Fixed incorrect handling of initial samples containing invalid values of discrete and categorical variables, which could negatively affect the Gradient-based optimization, Surrogate-based optimization and Adaptive design techniques.
  • Predictive modeling:
    • 19068 Fixed a bug in internal validation of models trained on samples with a grid-like structure (multigrid mesh), which could cause an access violation error in the case when the training sample contains a high number (order of \(10^5\)) of points.
  • NX block:
    • 19033 Fixed the inability to build the tree for a model containing dimension parameters of unsupported types.
  • DFBuilder block:
    • 19104 Fixed the GTDF/HFA/SurrogateModelType option not recognizing "GBRT" and "PLA" (actually supported approximation techniques) as its valid values.

pSeven 6.16.2

Release date: 7 February 2020.

This release includes pSeven Core (MACROS) 6.16.2.

Updates and Changes

  • Predictive modeling:
    • 18994 Added experimental support for exporting approximation models to the C# library format (.dll). Note that it requires a C# compiler installed, and both the source and binary C# export formats require an up to date license valid for pSeven 6.16 and above.
    • 18992 The automatic sample splitting algorithm, which is used in SmartSelection by default to split the training sample into the training and test subsets, was adjusted for the case of a large training sample (several thousands of points). This change helps to avoid certain negative effects on model quality, which were observed in previous versions when automatic splitting was used with large samples.
    • 18818 In the Model details dialog in Analyze, you can now export the model structure (for RSM and TBL models) and training log to text files — see Model Details.
    • 19003 Added a summary of model output properties to the end of the model training log. This summary lists training options applied to each output and contains a few other details, such as used approximation techniques and output dependencies and constraints, if they exist.
  • FloEFD block:
    • 19025 Added initial support for newer FloEFD versions (2019.x).
  • General:
    • 18875 Improved tooltips on the block information panel in Edit.
    • 18867 The text viewer in Workspace now allows viewing C# source code (.cs) files.


  • 18009 Updated section Model Details.
  • 19024 Fixed incorrect information about export formats supported for data fusion models in section Model Exchange and Approximation model block documentation.
  • 19004 Updated the 2.07 General - Dimension Reduction and 2.08 General - Dimension Reduction - Feature Extraction examples to use the more robust PCA dimension reduction technique.


  • Predictive modeling:
    • 19034 Fixed an issue where the trained model could show unexpectedly high errors for some outputs, if the search for linear dependencies between outputs is enabled.
    • 18999 Fixed incorrect preparation of the training sample in SmartSelection when model validation type is set to internal validation.
  • Uncertainty quantification block:
    • 18823 Fixed an issue with the manual configuration mode, where re-opening the block’s configuration dialog reset all options specified manually to their default values used in the SmartSelection mode, even if SmartSelection is disabled in configuration.
    • 18849 Fixed the convergence plot in the Reliability analysis (Sampling) technique report, which always showed confidence interval bounds for confidence level 0.95, regardless of the confidence level specified in technique configuration.

pSeven 6.16.1

Release date: 27 December 2019.

This release includes pSeven Core (MACROS) 6.16.1.

Updates and Changes

  • Predictive modeling:
    • 18509 Model export to C# source code is now supported for all approximation techniques (not yet supported for data fusion models, .gtdf). Note that C# export requires an up to date license valid for pSeven 6.16 and above.
    • 18962 Improved quality of approximation techniques based on Gaussian processes (GP, SGP and other) for noisy training data and for models with high input dimension.
  • Design space exploration block:
    • 18887 The Gradient-based optimization and Adaptive design techniques behavior in cases when the block cannot find a solution because of an infeasible problem or undefined responses is now consistent with the Surrogate-based optimization technique behavior. By default, mentioned techniques now output all evaluated designs and feasible designs (if any), leaving the optimal designs sample empty and allowing the workflow to continue. Also added an option which you can use to treat such cases as a block error and stop the workflow — see Unsolvable problem behavior in the Gradient-based optimization, Surrogate-based optimization, and Adaptive design technique options.


Compatibility Issues

  • This release changes the ways of handling of infeasible problems in the Design space exploration block, primarily concerning the Gradient-based optimization technique. Previously an infeasible problem caused an error, which you could handle in a few different ways (see Error Handling). Since 6.16.1, default behavior is to suppress the error and output the evaluated designs, if any. This may change workflow behavior if it contains conditional branching (Condition or a similar purpose block), that relies on the Design space exploration block finish status or result data. You may need to edit existing workflows or to set the Unsolvable problem behavior option to Technique error to reproduce old behavior.


  • Run:
    • 18963 Fixed pSeven not showing a message box with details on workflow issues when you attempt to run a workflow which contains validation errors or warnings.
    • 18968 Fixed an issue with block execution time statistics in long-running workflows, which produced error notifications and sometimes could crash pSeven.
  • Analyze:
    • 18803 Fixed an issue where pSeven sometimes could not acquire a license for some viewer and showed a placeholder instead of its actual content.
    • 18803 Fixed a rare error which lead to incorrect report behavior, identified by the following log message: Can't retrieve file id (can't locate ID).
  • Predictive modeling:
    • 18913 Fixed an error in the internal validation procedure, which occurred when the training sample contains many non-numeric values in the input part (values of variables), and the manual configuration mode option GTApprox/InputNanMode is set to ignore.
  • Design space exploration block:
    • 18895 18963 Fixed an issue where the block did not save the order of variables and responses correctly after some changes (for example, when you reorder variables, then delete one of them and create another with the same name). After this the block could output malformed results, as the bug made the order of variables or responses shown in the configuration dialog inconsistent with the column order in result samples. Note that while this bug is fixed in pSeven 6.16.1, a workflow created in prior versions of pSeven can contain a block saved with invalid configuration. When you open such a workflow in pSeven 6.16.1, it shows a validation error, which prohibits running the workflow. To remove the error, you will need to reopen the block, allow pSeven to resolve configuration issues, and verify the order of variables and responses afterwards. For details, see section Column Order in Results in Known Issues for the Design space exploration block.
    • 18938 Fixed an issue where the block could incorrectly evaluate linear and quadratic constraints when solving an optimization problem with discrete variables.
    • 18918 Fixed an issue with occasionally long solving time of the Surrogate-based optimization technique in problems with a high number of linear constraints.
    • 18931 Fixed an issue where the block could write incorrect values of generic constraints to result samples when solving an optimization problem with 2 or more constraints.
    • 18937 Fixed an error in the local search algorithm of the Surrogate-based optimization technique due to which it could not solve a constrained optimization problem with discrete variables if local search is set up to activate on the first iteration (the Local search option is set to Forced and Local search activation delay is default, 0).
    • 18888 18935 Fixed incorrect handling of the Orthogonal array design technique options due to which the technique could not start, and fixed the default value of the Multistart iterations option.
  • Uncertainty quantification block:
    • 18895 18963 Fixed an issue with incorrect column order in result samples similar to the one described above for the Design space exploration block. For details, see section Column Order in Results in Known Issues for the Uncertainty quantification block.

pSeven 6.16

Release date: 6 November 2019.

This release includes pSeven Core (MACROS) 6.16.

New Features

  • Run:
    • 18283 You can now copy and paste values of inputs, outputs and parameters in Run — for example, to use data from Excel, a text file, a Sample viewer in a report, or to copy some workflow’s outputs to another workflow’s inputs.
  • Analyze:
    • 18270 In the Sample viewer, the b_common_copy_value button or the Ctrl C hotkey on the Data tab copy values from the selected row to the clipboard. Paste them to the Inputs or Parameters pane in Run to set workflow’s inputs or parameters. When possible, pSeven matches the names in Run with names of columns in the Sample viewer. See section Sample Viewer for more details.
    • 16989 Added the command to import a Composite block cache to a report (see Import Data from Composite Block Cache).
  • Predictive modeling:
    • 15117 Added the SmartSelection support to the ApproxBuilder block. In the SmartSelection training mode, the block automatically selects and configures an approximation technique to obtain the most accurate model. This mode is now default, and you can switch back to the manual configuration in the block’s dialog.
    • 17970 18545 You can now specify a limited input domain (input constraints) for approximation models trained in pSeven. A model with limited input domain is non-extrapolating: it fits the training sample data but returns NaN outputs when inputs do not satisfy the domain’s constraints. This feature is supported in the ApproxBuilder block and the Model builder in Analyze. In the manual training configuration mode, it is enabled with the GTApprox/InputDomainType option. In the SmartSelection mode, you can set the domain using the Advanced option… hint.
  • Design space exploration block:
    • 18543 Added an optional local search algorithm to the Surrogate-based optimization technique (enabled with the technique’s Local search option). This is an adaptive algorithm, which intelligently limits the search scope in optimization in order to adjust the technique’s complexity to the response evaluation time and reduce the overall time in solving.
  • Approximation model block:
    • 18325 18457 The block now outputs model information and validation data as a report in the HTML format, which you can view in Analyze or export to a file on disk.
  • PythonScript block:
    • 16161 You can now test your scripts before running the workflow (see Testing Scripts).

Updates and Changes

  • Edit:
    • 18406 18417 Improved the block autolinking feature to better match complex port names and also to suggest creating commonly used links between ports which do not match by name.
  • Run:
    • 18036 Enabled the expand and collapse actions on the Parameters pane, so it now shows a parameter tree similar to the trees on the Inputs and Outputs panes.
  • Analyze:
    • 18373 18498 Noticeable performance improvements in 2D plots which contain much data or many datasets.
    • 18375 18267 Increased performance of the Import data from file dialog when working with large Excel files. Also improved table header recognition, and added a few more usability tweaks.
    • 18339 When multiple data series are selected on the Data series pane, the Rename… command can be used to edit all their names at once (see Rename for details).
    • 18249 pSeven now remembers the order of data series in the report database based on their index and keeps this order for training and test samples created using the Split data… command.
    • 17824 When you remove contents of a data series using the Discard values command, pSeven also cleans up the report database to reduce its size.
  • Analyze, Sample viewer:
    • 18056 18415 Made the viewer more responsive by avoiding unnecessary background data processing — for example, when you switch to the Dependency tab or change settings there.
    • 16275 You can now copy correlation coefficients and p-values from the Correlations tab, as well as the error and feature importance data from the Dependency tab.
    • 18259 Changed the default number format on the Data tab.
    • 18270 Added the number format b_sample_viewer_set_format quick button to the viewer’s title bar.
    • 18436 Cell tooltips on the Data tab now always show values with maximum precision, regardless of the number format in the table.
    • 17474 The general sample statistics now takes into account string data when counting duplicates and missing values.
    • 16267 Improved handling of long string values.
    • 18436 Clarified various validation messages which show up in the Issues pane when there are problems in viewer’s configuration.
  • Predictive modeling:
    • 17881 Improved the model validation procedure used by SmartSelection to estimate quality of intermediate models in the case when you do not supply a separate test sample. The new method provides more accurate quality estimates, eventually resulting in more accurate final models.
    • 17443 Extended the set of the Gaussian Processes (GP) technique options tuned by SmartSelection, potentially allowing it to create more accurate GP models.
    • 18320 18460 18462 Several performance and stability improvements for the Mixture of Approximators (MoA) technique. In particular, increased performance when parallelization is enabled (see GTApprox/MaxParallel).
    • 18394 Added the initial support for periodic kernel to Gaussian processes-based techniques, see GTApprox/GPType.
    • 18490 Reworked internal validation hints in SmartSelection. The new hint named Do not store internal validation data can be used to remove the cross-validation data samples from the model, potentially reducing its size. Advanced cross-validation settings can now be specified using the Advanced option… hint. See section Training Modes for details.
    • 18337 18026 When you export an approximation model as an Excel document with linked DLL, the DLL names are now based on the name of the Excel file for more convenience. Note that since DLL names (hence, the Excel document name) are used in the VBA macros code, the document name you specify must contain only characters, which can be represented in the system locale’s encoding (see Language for non-Unicode programs in Windows’ language settings). For compatibility across different local versions of Windows, it is recommended to use English characters only.
    • 18306 Redesigned the model training, retraining and update dialogs for better usability (see Model Training, Retraining a Model, Updating a Model).
    • 18269 Increased performance of Model validator in Analyze as now it does not reload data when you hide or show models in the Models pane.
    • 17969 17968 17908 Added information about model’s input and output constraints to the model details dialog (see the Constraints tab).
    • 18276 If you open the model details dialog for a model which is currently training, you can now stop training from this dialog, using the Stop training button.
    • 18497 18518 Added experimental support for model export to C#. The C# code export format is currently supported by the HDA, RSM, and TBL techniques. Note that using this feature requires an updated license file.
  • Design space exploration block:
    • 18352 Added special ports which set the Constant property of variables, named <variable_name>.Constant. Adding such a port to workflow parameters allows switching the variable to constant from Run, without opening the block’s configuration dialog.
    • 17348 The block now allows you to set Number of designs in optimization to 1, so you can use the Surrogate-based optimization technique to extend the initial sample with just one “next best” design.
    • 18469 18494 18521 The Adaptive design and LHS techniques now by default perform design space normalization internally to make their behavior more stable. Normalization can be disabled with the Normalize option.
    • 18063 When you use an initial sample with the Adaptive design technique, the block will calculate response values which are missing from the initial sample and include them in results.
    • 18467 Improved quality of the Adaptive design technique in cases when response evaluation failures can occur.
    • 18583 The Initial sample evaluation technique no longer removes duplicates in the initial sample by default, so the output sample and input sample sizes are the same. Also added the Evaluate duplicates option to this technique, which specifies whether to evaluate each duplicate input or re-use the same evaluation result for duplicates.
    • 17933 Updated the Orthogonal array design technique. The new version is significantly faster and more convenient to use. One of its distinctive features is the ability to adjust to any sample size by generating a “nearly orthogonal” array.
    • 18023 16990 17770 18500 18463 18248 18784 Several improvements and corrections in optimization algorithms resulting in higher stability and performance for certain problem types, in particular problems with no objectives (constraint satisfaction) or without constraints (unconstrained problems), and problems where design variables have significantly different order of magnitude.
    • 18374 If you select a technique intended to work with discrete variables but there are only continuous variables in configuration, the block now shows a warning to explain how such techniques work with continuous variables.
  • ApproxBuilder block:
    • 18255 Enabled model export to source code formats.
    • 18290 When you select an initial model, files with the .gta extension are also recognized as pSeven models.
    • 18290 Summary information output to the info port became more readable. This port is no longer monitored by default, as it is often unused.
  • Approximation model block:
    • 18538 The block can now load and evaluate data fusion models trained by the DFBuilder block. This also allows exporting data fusion models to C source code formats supported by Approximation model.
    • 18419 The Responses.Reference port is now always shown in block’s configuration. It accepts a test sample required for model validation.
  • Uncertainty quantification block:
    • 18250 Added the ports to output PDF and CDF data.
    • 18456 The block can now automatically export the analysis report to a file. Use the Report path port to set the export path.
    • 18250 Added the multinomial distribution support (user-defined discrete distribution).
    • 18250 The additional ports which set properties of variables and responses are hidden from the Ports and parameters dialog by default. To show them, use the b_blconf_uq_p button on the dialog’s toolbar or the Ctrl H hotkey.
    • 18474 Plot legends in analysis reports now support Unicode.
    • 18528 The block is out of beta state; removed the beta tag.
  • Direct integration blocks:
    • 18324 18331 18333 18334 ANSYSWorkbench, Excel, NX, and STAR-CCM+ blocks are updated to support the latest versions of their respective CAD tools.
  • NX block:
    • 17873 Clarified the error message which appears when you load an assembly but the block is unable to find some part files referenced in the assembly. The complete list of the part files required to load an assembly is now printed to the pSeven console.
  • STAR-CCM+ block (beta):
    • 18397 The block now correctly handles the case when it is unable to work with a project file which lacks certain features required to generate a mesh.
  • Program block:
    • 16198 The block now logs more information about jobs from a job array submitted to a SLURM cluster.
  • Text block:
    • 18054 Improved validation for names of variables. Python keywords and reserved names are now explicitly forbidden.
  • DRBuilder block:
    • 18285 Clarified some port and option descriptions.
    • 18285 The info port is no longer monitored by default as it is often unused.
  • Echo block:
    • 17245 The block now pre-formats Dict values which it prints to the run log for better readability.
  • Formula block (beta):
    • 18413 The block now correctly handles the case when it cannot calculate the function or the derivative of a function for some input values. Previously this situation could result in a failure to run a workflow, now the block outputs NaN values instead.
    • 18272 Block’s ports now allow only scalar and vector data types.
    • 18272 The block no longer calculates derivatives by default. When needed, you can enable derivatives with the new option Calculate the derivatives.
  • General:
    • 18428 When you edit a value using the Edit value… dialog, the data type in the dialog is now automatically set to the type of this value, instead of always selecting the Undefined data type as default.
    • 18057 If an error occurs when you start pSeven, it now shows a message containing error details and the startup log.
    • 17773 Improved corrupted database handling: now fixing such a database is always required before you start working with the project. This prevents working with a damaged project, which could lead to more errors.
    • 18322 18380 Updated some open source components to their recent versions, see section Open Source Components for details.
    • 18256 Added timestamps to the pSeven console log.
    • 18388 18499 18022 Several minor usability improvements.



  • Edit:
    • 18416 Fixed a GUI error which sometimes appeared when you remove a hierarchical port in the Composite block configuration.
    • 18381 Fixed a GUI error which could appear when switching to a horizontal workflow layout.
  • Run:
    • 18274 Fixed an unclear error message shown when a workflow is stopped due to some required inputs missing.
  • Analyze:
    • 18273 Removed the incorrect warning about non-numeric and empty values in Sample viewer, which always appeared when you switch to the Correlations or Dependency tab.
    • 18415 Fixed a bug in Sample viewer due to which it did not update the list of outputs on the Dependency tab when you added or removed dimensions in viewer’s configuration.
    • 18339 Fixed a bug in the Sample viewer, due to which it did a selection reset when you rename data series in the report database, it the Sample viewer is linked with a Page viewer.
    • 18789 Fixed a bug on the Correlations tab due to which zooming of a selected dimension slice to a 2D plot could work incorrectly after switching the scatter plot matrix between the X vs X and X vs Y modes.
    • 18056 Fixed a GUI error in Sample viewer which sometimes appeared when switching the predictions mode on the Dependency tab.
    • 18477 Fixed the GTApprox/TensorFactors option missing in the TGP technique settings in the Model builder.
    • 18044 Fixed the inability to reset manual plot bounds in the Model explorer after removing all models from its configuration.
    • 18573 18577 Fixed Model explorer malfunctioning when viewing a model which was trained with the TBL technique or viewing a MoA model obtained by updating an initial TBL model.
    • 18420 Fixed incorrect filtering of NaN and infinity values when calculating model accuracy metrics in Model validator using the model’s training sample.
    • 18562 Fixed Model validator showing inconsistent error percentile values on the quantile plot compared to values listed in the errors table.
    • 18420 Fixed an issue with logarithmic axis scales on error scatter plots in Model validator.
    • 18795 Fixed a bug in the Model validator configuration dialog due to which it kept displaying default tab names instead of real model names until you make some changes in configuration.
    • 18810 Fixed a freeze in the model details dialog when viewing a large TBL model.
    • 18245 Fixed a bug in 2D plot, 3D plot, and Parallel coordinates, which resulted in the inability to set number formats or scales if you select multiple data sources.
    • 18267 Fixed a bug due to which changing name prefixes in the data import dialog did not change the respective column names on the Result tab.
    • 18470 Fixed a bug which in rare cases could crash pSeven when updating a report.
    • 18310 Fixed a bug which sometimes prevented creating a new report while another report is still loading.
    • 18339 Fixed incorrect behavior of selections on the Data series pane when you rename data series in a report containing a Sample viewer or Page viewer.
    • 16515 Fixed incorrect rendering of certain Unicode characters in names of data series.
    • 18288 Fixed a bug which resulted in printing redundant messages to the pSeven console when exporting data from a report to a CSV or Excel file.
    • 17802 Fixed a bug due to which incorrect error messages could appear in the console log on the debug level when exporting the approximation model as an Excel document with linked DLL.
    • 18058 Fixed a bug which occurred when removing an expanded record from the project database and resulted in console error message with the debug level.
    • 18455 Fixed a GUI error which sometimes occurred when renaming a model in the Models pane.
    • 18260 Fixed the non-escaped HTML in tooltips for viewer title bars.
  • Predictive modeling:
    • 18461 Fixed a cross-validation bug in SmartSelection, which could lead to an error when training a model with categorical variables and point weights without a separate test sample.
    • 18478 Fixed an error in the Mixture of Approximators (MoA) technique when it is used with an initial model and a training sample which has a grid-like structure (multigrid mesh).
    • 18556 Fixed a bug in the HDA technique due to which a HDA model could produce false predictions of NaN outputs when the NaN prediction feature is enabled.
    • 18404 Fixed the inability to update a GBRT model which uses output transformation.
    • 18806 Fixed the RSM, SPLT, TA, iTA, and PLA techniques ignoring the linear output dependency option (see GTApprox/DependentOutputs).
    • 18603 Fixed the inability to retrain a model which was trained with certain advanced options specified manually.
    • 18050 Fixed the incompatibility of exported model C code with 32-bit compilers.
    • 18348 Fixed a bug in model export to C due to which exported TA and iTA models with SPLT factors could return NaN outputs for points which are out of the region covered by the training sample.
    • 18545 Fixed the inability to load models trained with the TBL technique to the Approximation model block.
    • 18573 Fixed the inability to change settings when retraining a TBL model in Analyze.
  • Design space exploration block:
    • 18817 Fixed the Adaptive design technique performing excessive evaluation of quadratic constraints when generating a DoE with dimension higher than 10.
    • 18450 Fixed the support for integer variables in problems with no objectives (constraint satisfaction problems).
    • 18444 Fixed incorrect behavior of some DoE techniques when generating a DoE with “frozen” variables (with equal lower and upper bounds).
    • 17994 Fixed the batch evaluation mode support in the Adaptive design for GP technique.
    • 18515 Fixed a bug that in rare cases could crash pSeven after uplinking a port which sets the Value property of a variable.
  • NX block:
    • 17873 Fixed the block leaving an empty file on disk after a blocked attempt to export an assembly to a part file.
  • Creo block:
    • 18454 Fixed the inability to export an assembly to the ACIS (.sat), Parasolid (.x_t), IGES (.igs), and VDA (.vda) formats.
  • STAR-CCM+ block (beta):
    • 18358 Fixed a bug due to which the block could incorrectly parse the structure of a model with no parameters and geometry built using the integrated 3D-CAD modeler. For such models, it created an empty geometry node in the project tree, and selecting this empty node resulted in a GUI error.
    • 18512 Fixed a bug due to which the model tree contained boundary conditions which cannot be used correctly in simulation as the block changed their type.
    • 18263 Fixed a bug resulting in the inability to process those parameters of the STAR-CAD, which were created manually instead of exposing them.
    • 18590 Fixed an error when changing the STAR-CCM+ Turbo.Path to data option.
  • Kompas3D block:
    • 17224 Fixed incorrect handling of OEM-encoded characters in paths and filenames.
  • Direct integration blocks:
    • 18396 Fixed a minor bug in the Copy value command.
  • Program block:
    • 17480 Fixed a bug due to which the block sometimes ignored most of the system environment variables if you add an environment variable in block’s configuration.
    • 16198 Fixed incorrect job status assignment when submitting a job array to a SLURM cluster.
  • Text block:
    • 18054 Fixed a GUI error which appeared if you rename a variable and then view results of test operations for this variable.
  • DRBuilder block:
    • 18285 Fixed the block ignoring the error handling settings when you specify an invalid evaluation budget.
  • DRBuilder, DRCompressor blocks:
    • 18285 Fixed incorrect validation of values assigned to some input ports.
  • DFPlayer, DRCompressor, DRDecompressor blocks:
    • 17708 Fixed an incorrect validation message, which showed up in the Issues pane when the model path is not specified.
  • Formula block (beta):
    • 18440 Fixed incorrect processing of NaN input values.
  • General:
    • 18323 Fixed pSeven always returning the 0 exit code when you run a workflow from the command line, even if execution errors occur.
    • 18067 Fixed a bug due to which pSeven sometimes ignored the project lock, which allowed multiple pSeven instances to work with the same project and could result in conflicting changes to this project.
    • 18600 Fixed a GUI error which appeared when an invalid value is specified for a port or an option which set some file path.

pSeven 6.15.1

Release date: 4 June 2019.

This release includes pSeven Core (MACROS) 6.15.1.

New Features

  • 18278 More approximation techniques can now work with an initial model. In addition to the GBRT technique, which supports it since version 6.8, you can now specify an initial model for the following techniques:
    • Mixture of Approximators (MoA): can update the initial model with new data or improve model accuracy using the same data as in the initial training.
    • High-Dimensional Approximation with Gaussian Processes (HDAGP): can use an existing HDA model as a trend, allowing the approximator to skip the internal preliminary training step, which can provide noticeable savings in training time.

Updates and Changes

  • Analyze, Model builder:
    • 14017 Added several internal improvements to the Mixture of Approximators (MoA) technique, in many cases making it faster and more accurate. This update is also available in the ApproxBuilder block.
    • 18278 Improved quality of GBRT models trained with an initial model.



  • 18287 ApproxBuilder block, Model builder in Analyze: fixed a possible crash when using a big training sample with point weights specified and output transformation set to "auto".

pSeven 6.15

Release date: 1 May 2019.

This release includes pSeven Core (MACROS) 6.15.


17946 This release requires an updated license file. Licenses issued for the previous versions of pSeven are not compatible with pSeven 6.15 and above. For details on how to update your license, see section Updating a License.

Note that updated licenses use a slightly changed license count method. Licenses for pSeven 6.14.4 and earlier limited the number of projects which you can open at the same time, but allowed running any number of pSeven instances. For example, a license could allow 3 active projects and any number of pSeven instances, causing a certain confusion. New licenses simply limit the number of pSeven instances. This change does not affect your license terms: for example, if your current license allows 3 active projects, you will receive a new license which allows 3 pSeven instances.

New Features

  • Design space exploration block:

    • 17932 The block can now calculate and output design feasibility measures showing the degree of constraint violation. See section Results in the block documentation for details.
    • 17932 You can enable additional ports which output the index of current iteration, which can be used by blackbox blocks. See section Blackbox Ports for details.
    • An initial sample can now be sent to the block as two matrices containing values of all variables and all responses. See section Samples and Designs.
    • Similarly, the block can output results as two matrices containing variables and responses. See Samples and Designs for details.
    • Various improvements in the Adaptive design technique including:
      • 17955 Significantly increased performance when generating a high-dimensional DoE with no objectives but with a high number of linear constraints, solving the specific problem of generating a uniform DoE in a narrow subset of the design space in high dimension.
      • 15764 Improved technique performance and stability when generating a DoE with functional constraints.
      • 17935 18042 Better support for sample-based generation when response evaluation is disabled. The technique will now generate uniformly distributed points instead of concentrating them in a limited design space area.
    • 17885 The Surrogate-based optimization and Adaptive design techniques now fully support discrete variables with non-integer values.
    • 17869 The Adaptive design for GP technique now supports multi-dimensional responses.
  • Predictive modeling:

    • 17914 17723 Added a special model training mode, in which pSeven tests the training sample to find linear dependencies between responses, and trains a model which keeps these dependencies. This mode is available:

      • In Analyze, when using the Model builder tool in the SmartSelection mode (the Dependent outputs hint) or the manual mode (the Advanced/DependentOutputs option).
      • In the ApproxBuilder block, using the GTApprox/DependentOutputs option.

      For details on the new method, see section Output Dependency Modes in the pSeven Core manual.

    • 17914 17724 The SmartSelection mode in the Model builder tool is updated with a new, even more accurate and automatic method which decides whether to apply log transformation to the outputs data in the training sample. Log transformation may improve model accuracy when training outputs are exponentially distributed. To enable the new method, use the Try output transformations hint.

    • 17886 18000 17989 Improved quality and stability of all approximation techniques based on Gaussian processes. In particular, the robust version of the Gaussian processes algorithm was significantly improved and now provides both high accuracy and stability. This version of the algorithm is now default in all modeling tools.

    • 17846 Improved load balancing and computational resource utilization when training models in parallel mode.

    • 17959 Overall improvement in quality and performance of the SmartSelection mode thanks to a number of updates in approximation techniques used by this mode.

    • 17726 You can now remove the accuracy evaluation information from approximation models when exporting the models to different formats which are not specific to pSeven. Removing these model sections can noticeably reduce size of the exported model. Note that when you export a model to a pSeven binary model format (.gtapprox), you can also remove smoothing data, training sample, internal validation data, annotations and training log sections.

  • 17730 The UQ block was redesigned and updated, also renaming it to Uncertainty quantification. Key features of the updated block include:

    • Support for sensitivity and reliability analysis studies using several methods.
    • Results output as HTML-formatted reports which you can load into a Page viewer in Analyze.
    • Redesigned configuration dialog similar to the Design space exploration block.
    • SmartSelection support: the block automatically selects the most suitable analysis method based on your configuratiuon of variables and responses.
    • Improved analysis quality and performance.
    • Support for advanced configuration scenarios including vector variables and responses, setting their properties dynamically using special input ports, batch evaluation mode allowing parallelization, and other.
  • Excel block:

    • 17834 The block now supports paths to the Excel document and export file as inputs sent to the new Document file and Export file ports. This means that you can dynamically load different documents to the block during a workflow run, or select the new ports as workflow parameters so the document paths can be quickly changed in Run. See the Excel block documentation for details.
  • FMI model block:

    • 17458 The block is promoted to release blocks and is no longer tagged as beta.
    • 17458 Added Model Exchange support: the block can now integrate Functional Mock-up Units for Model Exchange, supports binary FMU formats and can handle 32-bit FMI models.
  • STAR-CCM+ block (beta):

  • Edit:

    • 17892 You can now set custom icons for the Composite, Program, PythonScript, and Text blocks. See Block Icons for details.
  • Analyze:

    • 17966 The Design space exploration, Approximation model, and Uncertainty quantification blocks received short default names for convenience (DSE, Model and UQ). These names are used in workflows and in the project database, making navigation in the project database a bit easier.
    • 17732 Added the linked selection support to the Page viewer. You can use a Page viewer in the linked selection mode to show additional details for rows you select in a linked Sample viewer or points on a linked 2D plot or Parallel coordinates plot. See section Linked Selection for details.
    • 17737 When importing data to a pSeven report, you now can switch to the data append mode to update existing data series if their names match the names of columns in the imported CSV or Excel file. The append mode is enabled on the Result settings pane of the Import data from file dialog.

Updates and Changes

  • Workspace:
    • 17440 Improved the project locking mechanism in order to prevent situations when the same project is open by two or more pSeven instances — for example, when a project is stored on a network drive or in another location with shared access. Now its is also possible to force unlock a project on open.
  • Run:
    • 17344 Now when you reset a parameter value in Run, this value is removed from the parameter preset and the default port value specified in block configuration is used instead.
  • Analyze:
    • Data import dialogs:
      • 17962 Added the ability to clear prefixes for the names of columns imported from a CSV or Excel file using the b_common_sweep button in the Result settings pane.
    • Sample viewer:
      • 12685 Now you can export the plots from the Correlations and Dependency tabs to a file using the Export image dialog where you can select the image size.
      • 17852 17904 Improved tooltips for the feature importance plot on the Dependency tab to make them more informative. The tooltips are always shown when hovering the bars and their labels.
      • 17718 Simplified value representation in table cells, the original value format is used only when copying cell values to the pSeven clipboard.
      • 17491 Increased the viewer performance when working with large datasets.
      • 17491 Various minor visual fixes and improvements.
    • 2D plot:
      • 17445 Now it is possible to save copies of the data currently shown on the 2D plot to the report database with the help of the Create new data series… command. It respects all selections and filters applied in the viewer’s configuration: new data series added to the report database contain only those values from the original data series which pass the filters.
    • Page viewer:
      • 17997 The viewer behavior is changed in cases when there is only one data series in the viewer’s configuration — no table or placeholder is displayed.
    • Text viewer:
      • 17789 Now the b_text_viewer_edit_mode button is disabled when the viewer is minimized.
    • Predictive modeling tools:
      • 17818 Now model training time is shown in model details and in the Model pane in Model validator.
      • 17796 You can extract the training sample, if a model stores the data used in its training, with the help of the b_extract_sample button on the Training sample tab of the model details dialog.
      • 17487 Improved the support for categorical variables in the SPLT and GBRT approximation techniques.
      • 17842 When loading models trained in versions prior to 6.8, pSeven now automatically calculates the \(R^2\) error metric which these old versions did not store to the model, and adds it to model details.
      • 17498 Improved representation of outputs approximation techniques on the Parameters tab in the model details dialog.
      • 17839 Now it is possible to export model details to plain text using the Export model details to file button on the Parameters tab. Note that for big models (with many variables or submodels), only the export to a file is available and copying model details to the pSeven clipboard is disabled.
      • 17858 Clarified the error message shown when the export of a model to Excel or executable failed with an error.
    • Model explorer:
      • 17806 When you select which model inputs and outputs to plot, the viewer now shows their full names in the selection lists, instead of always using the default names.
    • General:
      • 17850 You can now hold the Shift key and use the mouse wheel for horizontal scrolling in reports.
      • 17750 17727 Added the DUPLEX method to split data samples into train and test subsets. See Split Data for details.
      • 17903 CSV files are now exported with headers by default.
      • 17859 Improved stability of various project and report database management operations.
  • Issues pane:
    • 17884 Added a close button.
    • 17884 Validation warnings shown on the pane can now be copied to the pSeven clipboard using the Ctrl C hotkey.
  • Block configuration:
    • 17833 Now the block’s type is not shown on the block and in the header of the configuration dialog if such a type is the same as the block’s name.
    • 17426 Improved validation for monitored ports, in particular:
      • A validation warning is shown in cases when you enabled port monitoring for blocks nested in a Composite block with enabled parallel execution.
      • A validation error is shown in cases when the automatically generated alias of a monitored port is not unique.
  • Model builder, ApproxBuilder block:
    • 18000 Added the Auto setting for the GTApprox/GPLearningMode option which selects different versions of the training algorithm for the Gaussian processes-based techniques. This setting is now default.
  • ApproxBuilder block:
    • 17910 Added the ability to export approximation models to different formats which are not specific to pSeven. See the ApproxBuilder block documentation for details.
  • Composite block:
    • 17797 Ports tab improvements:
      • Enabled batch operations for adding ports.
      • Disabled monitoring by default for do and done ports after uplink.
      • Added the ability to remove ports using the Delete button.
  • Design space exploration block:
    • 15764 Significantly improved performance of several internal algorithms used in constrained optimization problems.
    • 17817 17821 17922 Improved stability in mixed-integer linear problems which include variables bound to a very small range (order of \(10^{-6}\) and less).
    • 16693 Noticeably reduced memory consumption for some types of optimization problems.
    • 17932 Empty matrix and a single NaN value are now supported as valid blackbox response values indicating that evaluation failed. Previously, the block always required the correct shape of the response matrix.
    • 17932 Changed the default value of Latin hypercube sampling.Iterations option to 20.
    • 17934 Added validation for cases when the blackbox parallelization ratio is greater than 1 and the batch mode is disabled in run options.
    • 17934 Added a new Computational scalability option for the Adaptive design for GP technique.
    • 17934 Updated the Train iterations option description for the Adaptive design for GP technique.
  • Excel block:
    • 17919 You can clear the loaded document file and the export file to save changes using the b_common_sweep on the Document file and Export file panes.
    • 17834 Improved validation when creating new ports.
  • FMI model, STAR-CCM+ (beta) blocks:
    • 17458 17899 Now you can manually interrupt the block’s execution when a workflow is running. The block will not try to output intermediate results.
  • FMI model block:
    • 17458 The model tree now shows local variables of the enumeration type which can be changed and read in the workflow.
    • 17458 Added validation of the Tree node property in the Variables pane in block configuration.
  • STAR-CCM+ block (beta):
    • 17915 The model tree now shows global parameters and plots presented in the simulation.
    • 17915 Clarified an error message shown when the block tries to update model geometry.
  • GUI:
    • 17697 The Distribution block is now deprecated and removed from the Block Library.
  • General:
    • 17793 pSeven now does not use system proxy settings for local connections. If a connection error occurs during pSeven startup, it shows an error message describing the issue and possible ways to resolve it.
    • 17056 Better file handling prevents file access errors on Linux when pSeven starts new processes such as external solvers and other tools.
    • 17337 Increased responsiveness of the user interface when updating data series using the report refresh command.
    • 17751 Some performance improvements when loading and saving workflows and reports.
    • 17337 Minor visual improvements in the report database, in particular added tooltips for the data source column.


Compatibility Issues

  • In this release, the Uncertainty quantification block was upgraded to replace the old UQ block. When you open a workflow which contains a UQ block in pSeven 6.15, this block is automatically upgraded to Uncertainty quantification. Note that workflows containing the old UQ block created in versions earlier than pSeven 6.9 are not upgraded automatically. For versions below 6.9, a two-step upgrade can be performed: first to pSeven 6.9 and then to 6.15.
  • Prior to 6.15, it was possible to train an approximation model with GTApprox/GPLearningMode set to Robust and GTApprox/ExactFitRequired set to True. This usually produced a bad quality model, as the robust Gaussian processes algorithm in fact does not support exact fit. Since 6.15 this combination of options is explicitly prohibited and raises an error before training.
  • The --runner command line option is no longer used to launch pSeven Runner. This mode is now selected by specifying the -- no edit --no-analyze options. See section pSeven Runner for details.


  • Workspace:
    • 17440 Fixed the bug due to which it was possible to upgrade a project opened by another pSeven instance.
  • Edit:
    • 17930 Fixed incorrect behavior when trying to uplink two ports with incompatible types to the same port.
    • 17749 Fixed the bug due to which the link attached to a port could disappear when recreating this port.
    • 17478 Ports with complex types, like the Number of designs port in the Design space exploration block, are now uplinked properly: new ports created by uplinking will have the same types.
    • 17893 Fixed the inability to switch the value type for a File port.
  • Run:
    • 17825 Fixed the bug resulting in workflows not being marked as changed after clearing run statistics using the Clear statistics… command.
    • 16252 Fixed a GUI freeze when switching to the Statistics tab.
  • Run, Analyze:
    • 18236 Fixed a rare bug which resulted in pSeven crash when refreshing the report if the data to update was received from the currently running workflow.
  • Analyze:
    • Sample viewer:
      • 17445 Fixed incorrect behavior of the Create new dataseries… command that does not respect filter settings applied in the viewer’s configuration.
      • 17918 Fixed the bug due to which changing the number format for selected data sources may result in changing formats of other data sources.
      • 17918 Fixed the bug resulting in the inability to set number formats to data sources after removing some dimensions in the viewer’s configuration.
      • 17792 Fixed the Correlations tab sometimes not correctly showing axis labels of scatter plots in the X vs Y display mode when the viewer is maximized.
    • 2D plot:
      • 17957 Fixed a rare bug which may occur if there are datasets of various dimensions and also extrapolation and error type enabled in the viewer’s configuration.
      • 17788 Fixed incorrect behavior of Copy from and Copy to commands not showing new datasets when they were added to the current configuration of the viewer.
      • 17788 Fixed the bug with a dataset name not being reset to a default one after it was removed from the viewer’s configuration.
      • 17794 Fixed incorrect behavior of zoom function in case of extrapolated points.
    • Parallel coordinates plot:
      • 17925 Fixed a misleading warning message related to improper dataset configuration.
      • 17963 Fixed a few minor bugs that occurred when switching between datasets on the plot.
      • 17001 Fixed the issue with plots being sometimes displayed incorrectly in case of high-dimensional datasets.
      • 17001 Fixed incorrect behavior of blending dataset setting.
    • Page viewer:
      • 17683 Fixed the error which occurred when switching between datasets using the up and down arrow keys.
    • Text viewer:
      • 17789 Fixed incorrect behavior of the Duplicate command that does not respect the minimized state of the original viewer.
      • 17789 Fixed the mode settings being reset after duplicating the viewer.
    • Predictive modeling tools:
      • 17038 Fixed various usability issues in predictive modeling tools, related to handling models with categorical variables.
      • 17936 Fixed the bug that could sometimes lead to inability to save a report in pSeven on Windows if the model training in the report was interrupted or stopped with an error.
      • 17858 Fixed a rare bug resulting in the inability to export an approximation model as an Excel document with a linked DLL.
      • 17866 Fixed the bug due to which the Retrain command could work incorrectly for smoothed models.
      • 17359 Fixed a bug due to which an original model was sometimes replaced after retrain.
      • 17359 Fixed incorrect behavior of the Retrain command that does not respect internal approximator options.
      • 17839 Fixed the bug due to which it was impossible to copy model details of big models to the pSeven clipboard.
      • 17839 Fixed the bug due to which it was possible to copy model details to the pSeven clipboard before the detailed information about a model is shown in Model details dialog.
      • 17839 Fixed the bug resulting in improper calculation of the default number of points per dimension for the Full factorial design technique in the Make predictions dialog for models with many outputs.
      • 17967 Fixed a bug in SmartSelection which caused a crash with a training sample containing categorical variables, when the sample is split into train and test subsets, and internal validation is disabled.
    • Model validator:
      • 17847 Fixed axis tick labels overlapping in quantile plots.
    • Model explorer:
      • 17482 Fixed the issue with the same slice settings option not being applied on the plots even if it was enabled in the viewer’s configuration.
      • 17482 Fixed various minor bugs and inconsistencies.
    • GUI:
      • 17491 Fixed a slowdown when opening a Split data dialog with a large number of data series in the report.
    • General:
      • 17850 Fixed automatic scrolling not working when moving a tile to the top or bottom of the report.
      • 17747 Fixed a bug that occurred in some rare cases when renaming or removing records in the project database.
      • 17880 Fixed a bug in the CART sample splitting algorithm due to which it could produce unacceptable split results when working with a big sample which contains too many duplicates.
      • 17998 Fixed a bug due to which export to Excel created a document with Cyrillic characters in sheet names.
  • Design space exploration block:
    • 18042 17375 Fixed a bug in processing response evaluation results in specific cases when configuration includes both generic and cheap responses. When reaching the evaluation limit for a generic response, the block stopped evaluating the response, but further internally treated new points where the response was not evaluated as points where it failed to evaluate. This did not affect the solution quality, but the skipped points were incorrectly attributed as infeasible, which led to misinterpretation of results.
    • 18042 NaN values of responses in initial samples are now processed as special values indicating that a response failed to evaluate. Previously such points were simply excluded from consideration.
    • 15711 Fixed a minor bug in the surrogate-based multistart method due to which it behavior depended on the Computational scalability option while it should not.
    • 17376 17901 Fixed the support for discrete variables in the Adaptive design technique.
    • 17901 Fixed the Adaptive design technique sometimes ignoring the initial samples.
    • 17916 Fixed the Adaptive Design technique crashing when constraints in the generation space are defined in such a way that there is only one feasible point.
    • 17441 Fixed the Orthogonal Array technique ignoring the Deterministic option.
  • ApproxBuilder block:
  • Excel block:
    • 17906 Fixed an error on loading an Excel document with macros if VBA trust access is disabled in Excel settings (trust access is not required to load a document, only to get names of macros automatically).
  • Program block:
    • 17059 Fixed a bug due to which the block could incorrectly generate a submit script when connecting to a Torque cluster if the number of compute nodes to allocate for the job was set to 0 (default).
  • GUI:
    • 17883 17853 Fixed GUI errors from 3D plots in Analyze which appeared when the plot cannot be rendered properly.
    • 17864 Fixed a bug with overlapping viewers in reports where many viewers are open.
    • 17887 Fixed the toolbar disappearing in some viewers after maximizing.
  • General:
    • 17832 Fixed a rare error which occurred when uplinking any block’s port to the port of a workflow root block (workflow inputs) having a default value if the ports are of incompatible types and mutual value format conversions are impossible.
    • 17745 Added missing validation of inputs, which are received during a workflow run, to ports that have complex types with additional restrictions on valid values. Previously input values for such ports were validated only when editing the workflow, and if an invalid value was received from some block in a running workflow, it could cause a crash instead of stopping the workflow with a meaningful error.

pSeven 6.14.4

Release date: 1 February 2019.

This release includes pSeven Core (MACROS) 6.14.4.


  • Design space exploration block:
    • 17863 Fixed a bug in pSeven for Windows due to which the Adaptive design technique and optimization techniques could crash the block in certain configurations because a memory access violation error occurred.
  • STAR-CCM+ block (beta):
    • 17882 Fixed a bug in handling the inlet boundary conditions: the block never changed values of velocity magnitude and total pressure before running a STAR-CCM+ simulation, so it produced irrelevant results when these boundary conditions were used in the workflow.
    • 17893 Fixed the inability to build the project tree for a simulation containing boundary conditions of unsupported types.
    • 17893 Fixed the error which occurred when simulation includes a user field function of the Vector type which references other field functions.
  • ApproxBuilder block:
    • 17500 Fixed a bug in the heteroscedastic Gaussian processes algorithm due to which the GTApprox/Heteroscedastic option had no effect.
    • 17500 17468 Fixed incorrect handling of NaN values in the noise sample (the output_noise_variance port) due to which some valid samples containing NaN values were not accepted by the block.

pSeven 6.14.3

Release date: 13 December 2018.

This release includes pSeven Core (MACROS) 6.14.3.

New Features

  • 17739 Introduced a new license feature, pSeven Runner and Post-Processing. This is a separate launch mode similar to pSeven Runner: it disables Edit but allows running existing workflows and processing their results in Analyze. It is intended for users who work with run-ready workflows from pSeven GUI and need the Analyze tools for interactive data analysis and visualization, while pSeven Runner is intended for using pSeven in batch mode (see Command-Line Workflow Integration).

Updates and Changes

  • Edit:

    • 17342 The background grid is no longer shown when you move the workflow graph. It has been found that rendering the grid while moving causes a noticeable slowdown on some systems.
  • Analyze:

    • 2D plot:
      • 17710 Added tooltips for extrapolated points and removed the warning which appeared when hovering such points.
      • 17710 A few minor improvements in plot visuals.
    • Predictive modeling tools:
      • 17783 Added the support for exporting an approximation model as a Functional Mock-up Unit for Model Exchange (FMI standard).
    • General:
      • 17710 The Issues pane now shows warnings for plots with a logarithmic scale axis, if there are points which cannot be displayed because their coordinates for this axis are 0 or negative.
      • 17740 Added tooltips in the Split data dialog.
  • Optimizer block:

    • 17506 Optimizer is now a legacy block; it is recommended to use the Design space exploration block instead. The Optimizer block will no longer receive any updates and can be removed in future pSeven versions.

      Note that Optimizer is still available although it is hidden from the Block library pane by default. To show it, enable legacy blocks in the Block library pane’s context menu b_context.

  • Design space exploration block:

    • 17755 Changed manual stop behavior to be consistent with other blocks which support this feature. The Stop button is now always enabled in the dialog; previously the block could disable it when intermediate results were not available, which made the dialog behavior confusing.
    • 17678 Slightly improved speed and quality of the Adaptive Design technique.
  • Approximation model block:

    • 17783 Added the support for exporting a model as a Functional Mock-up Unit for Model Exchange (FMI standard).
  • Excel block:

    • 17774 You can now specify the value of the Macro name option manually when editing it on the block’s Options tab or on the parameters pane in Run. In previous pSeven versions, the macro to run could only be selected from the list of macros detected by the block on document load, so the option was unusable as a workflow parameter if the block did not contain a preloaded file.
  • FloEFD block:

    • 17752 Added automatic handling of the error which occurs when FloEFD for NX fails to close NX correctly. The block now detects this situation and automatically terminates all NX processes with a 10 second timeout if the error occurs. See FloEFD for NX Error on Exit for details.



  • Run:

    • 17712 Fixed the bug which could freeze pSeven when editing a parameter which is mapped both to a StringScalar port and a special port which receives a file path string.
  • Analyze:

    • Parallel coordinates plot:
      • 17681 17826 Fixed a few minor bugs due to which linked selection in parallel coordinate plots sometimes did not work as intended.
    • Model validator:
      • 17756 Fixed incorrect calculation of the \(R^2\) metric when internal validation data is selected as a reference sample.
    • General:
      • 17703 Fixed a warning which appeared when enabling linked selection in a report created in previous versions of pSeven, if this report contained viewers which were created using the Duplicate command.
      • 17713 Fixed the report refresh command incorrectly updating data series which use the “<last run>” record as their data source, if the refresh is done before new data is generated by the workflow.
  • Design space exploration block:

    • 17813 Removed the invalid No threshold value from the Convergence threshold for gradients option of the Gradient-based optimization technique. In fact this threshold cannot be disabled in gradient optimization methods, and this setting never worked in such a way.
    • 17813 Fixed the Number of designs setting tooltip showing an incorrect number of designs to be generated when the Full factorial design technique is selected.
  • Approximation model block:

    • 17714 Fixed the slowdown which happened when saving or loading workflows containing an Approximation model block with a big model loaded into its configuration.
    • 17714 Fixed the model tree not being loaded automatically when you specify the model by assigning a value to the Model file or Model path port.
  • Excel block:

    • 17767 Fixed the inability to load a document file where only one cell is filled if this cell contains a string.

    • 17774 Fixed incorrect macro detection: the block assumed that all macros are located in the default module (Module1) of the VBA project. As a result, macros from other modules were not detected, and names of detected macros did not include the module name. The latter lead to errors when testing macros in the block configuration dialog.

      Note that to correct this error in workflows created using previous versions of pSeven, you will need to reload the document in the Excel block (click b_xlsload on the document toolbar).

  • NX block:

    • 17685 Fixed incorrect handling of NX assemblies by the block which caused the “File not found” error message when opening a ZIP archive with the assembly.
  • PythonScript block:

    • 17741 Fixed incorrect automatic resize of the Variables pane after maximizing the block configuration dialog.
  • Text block:

    • 17776 Fixed quick selection sometimes not detecting field delimiters correctly.
  • FMI model block:

    • 17677 Fixed incorrect block behavior in the case when the path to the user or system temporary directory contains local language characters.
  • GUI:

    • 17762 HTML descriptions of ports and parameters are now additionally validated, so an occasional invalid HTML code in a description does not corrupt the rendering of workflow inputs, outputs, and parameters in Run.
  • General:

    • 17413 Fixed pSeven resetting the value of a parameter specified in Run when the corresponding port is renamed in block configuration in Edit.
    • 17728 Fixed a bug in handling a corrupted project database which could cause a slowdown or freeze during database repair.

pSeven 6.14.2

Release date: 14 November 2018.

This release includes pSeven Core (MACROS) 6.14.2.

Updates and Changes

  • 17704 STAR-CCM+ block (beta): added the support for Power-on-Demand licenses.



  • 17704 STAR-CCM+ block (beta): fixed a bug in connection to a remote STAR-CCM+ server which caused the error message: “Document path is not set.” when trying to connect.

pSeven 6.14.1

Release date: 31 October 2018.

This release includes pSeven Core (MACROS) 6.14.1.


17721 Since this release, the updates of pSeven and pSeven Core based on a recently released version will have short version numbers, for example 6.14.1 instead of 6.14 Service Pack 1.

Updates and Changes

  • 17687 Design space exploration block: added the button to show or hide disabled ports to the Ports and parameters dialog toolbar. Previously this command was available only from the context menu or using the Ctrl H hotkey.
  • 17707 FloEFD block:
    • Added initial support for using the block with FloEFD 18. Note that this update is based on the not yet released FloEFD 18 Beta for NX and is provided as a preview only.
    • Added correct handling of errors caused by unsupported characters in names of FloEFD goals.


  • 17707 Added a note on unsupported goal names in FloEFD projects — see section Known Issues on the FloEFD page.


  • 17705 Analyze: fixed a bug in the Parallel coordinates plot viewer which caused instability when loading or editing reports containing these plots and could sometimes crash pSeven in Analyze.
  • 17687 Design space exploration block: fixed incorrect error messages shown when the block discovers an infeasible problem — for example, a problem where constraints cannot be satisfied.
  • pSeven Batch:
    • 17709 Fixed a bug due to which the --runner command line option was required to use a Runner license (see pSeven Runner) when you start pSeven in the batch mode. pSeven Batch will now automatically use a Runner license when it is available.
    • 17435 Fixed a possible crash on exit from pSeven Batch under Windows. The crash happened after a workflow run finishes, and did not affect the results.

pSeven 6.14

Release date: 11 October 2018.

This release includes pSeven Core (MACROS) 6.14.


17363 This release requires an updated license file. Licenses issued for the previous versions of pSeven are not compatible with pSeven 6.14 and above.

For more details on how to obtain a new license file, see section License Setup.

New Features

  • 9999 16153 11821 17268 The Design of experiments block is upgraded to Design space exploration with the following key features:

    • The entire set of optimization, adaptive and design of experiment techniques under the hood. You can easily combine strategies and switch between techniques when solving your design tasks.
    • Four new tecnhiques implemented: Surrogate-based optimization, Gradient-based optimization, Adaptive design for GP, and Diagonal sampling.
    • Added support for discrete and categorical variables.
    • Added support for adaptive generation of designs based only on the initial sample.
    • Smart selection technology which automatically selects the best approach based on the properties of your design task.
    • Redesigned interface with improved responsiveness and usability.

    When you open a workflow which contains a Design of experiments block in pSeven 6.14, this block is automatically upgraded to a Design space exploration block.

  • 17331 Added the FMI model block (beta). This block enables integration of Functional Mock-up Units for Co-Simulation into pSeven workflows.

  • 16104 Added the STAR-CCM+ block (beta). This block which enables integration of STAR-CCM+ simulations into a workflow, supports CFD analysis in STAR-CCM+ and can run simulations on a remote STAR-CCM+ server.

  • Analyze:

    • 17275 Added connected interactivity between Sample viewer, 2D plot and Parallel coordinates plot via the linked selection option. You can now synchronize selections between these viewers if they share at least one common data source and also have linked selection (toggled with the b_linked_selection button) enabled. Actually it means that selections you make in a viewer also select corresponding data in other viewers linked to it. See section Linked Selection for more details.
    • 17074 Significantly increased Model builder performance in the SmartSelection mode thanks to the new method of model training parallelization which utilizes multi-core CPUs in a much more efficient way, training candidate models in parallel. This feature is also supported in the manual training mode in Model builder and in the ApproxBuilder block, where it can be controlled using the GTApprox/SubmodelTraining option. For underlying details, see section Submodels and Parallel Training in the pSeven Core manual.
    • 15107 17326 SmartSelection in the Model builder can now build more accurate models in cases when values of some training outputs are exponentially distributed, thanks to the improved support for automatic log transformation of training data. You can also specify output transformations manually on the Data settings pane in Model builder.
    • 15107 Now you can add custom settings in the SmartSelection mode, for example, manually select the approximation technique, limit the number of parallel threads or set any arbitrary option used in model training. You can configure specific training options and set their values via the Advanced option hint in the Model builder tool. See the updated Training Modes section for details.
    • 15449 15542 Added the Variables tab to model details dialog. This tab contains metainformation about model’s inputs and outputs. If names of variables were specified when training the model, these names will now be used in model details instead of default ones. Added the ability to copy model details to the pSeven clipboard, now it is also possible to copy RSM models’ terms as formulas using the b_copy_as_formula button.
  • Text block:

    • 16562 You can now quickly view the results of your testing operations — values of variables read from the template and a preview of changes applied by the block to the template. Use the b_locate button to scan changes in the template. See section Locating Variables for details.

Updates and Changes

  • Edit:
    • 16988 Simplified port naming when grouping some blocks to Composite.
  • Edit, Run, Analyze:
    • 17244 Improved responsiveness of the user interface when scrolling tabs. Now navigation between different open workflows and reports is noticeably faster.
  • Analyze:
    • 13286 Added a command to navigate to a specific row in the Sample viewer (triggered with the Ctrl G hotkey).
    • Sample viewer, Parallel coordinates plot:
      • 13286 17210 Increased viewer performance when working with large datasets.
    • 2D plot:
      • 17465 Increased the default display limit in the viewer configuration. Now up to 10,000 points are shown for each dataset.
      • 17676 Improved the display of images shown in tooltips for the points on the plot.
    • Column chart:
      • 17277 Added horizontal scrolling for large column charts that is enabled in viewer configuration.
    • Models pane:
      • 17230 Models containing a big training sample now load noticeably faster.
    • Model explorer:
      • 16580 Added the ability not to display sensitivity scores on plots. They are now disabled in the viewer configuration.
      • 16580 17233 Minor visual improvements. In particular, improved tooltips on the Models pane.
    • Model builder:
      • 16949 16951 You can now easily stop training of the GP, HDAGP or SGP model with an intermediate result. The time to finalize a model is significantly reduced. If the model training is terminated, accuracy evaluation for the model is not available even if it is requested.
      • 16175 Improved the special data analysis algorithm used in SmartSelection to prevent oscillations of model outputs when a smooth model is required.
    • General:
      • 17326 Updated the CART algorithm used to split data samples (see Split Data) in order to better handle samples with duplicated points.
      • 17259 Improved corrupted database handling: now you can remove damaged entries and optionally save a backup copy of your corrupted database.
      • 16941 Added the ability to set name prefixes to data series when importing data from CSV or Excel files.
      • 16159 Revamped Import data from CSV and Export data to CSV dialogs. Now the resized dialogs have a better layout with improved option names and settings.
  • Optimizer block:
    • 17012 When provided with an initial data sample, the block now does not request additional evaluations for linear and quadratic functions but restores their analytical form using the initial sample data only.
    • 17476 Improved multi-objective optimization algorithms for high-dimensional problems.
    • 16688 Some performance improvements in surrogate-based optimization methods.
  • Model builder, ApproxBuilder block:
    • 17074 Added the "Auto" setting for the GTApprox/RSMStepwiseFit/inmodel option which selects the type of initial RSM model when using the stepwise regression method. This setting (now default) performs automatic selection based on the number of model terms. This modification noticeably speeds up stepwise regression RSM model training in some cases, for example, in case of quadratic dependency with multiple inputs or in presence of categorical variables.
    • 17367 Added new special values for the HDA technique options which enable automatic tuning for them. These values are now default. Updated options are:
    • 16951 The maximum training sample size supported by the GP and HDAGP techniques is now limited to 4000 points.
    • 16951 Added an upper limit of 4000 for the GTApprox/SGPNumberOfBasePoints option.
    • 17419 Optimized the exported C code for the accuracy evaluation (AE) information in Gaussian Processes models. Compiled GP models with AE are now up to 2 times smaller. This change does not affect the models which do not store AE information.
  • ApproxBuilder block:
    • 17362 15542 Added sample_descriptions, sample_quantities, and sample_units ports to provide more details on variables and responses of the approximation model trained by the block. This information will be saved to the model and can be easily retrieved when you view the model in Analyze or load it into the Approximation model block. Now it is also prohibited to assign non-unique names to model inputs and outputs received to sample_names port. See the updated ApproxBuilder documentation for full details.
  • Approximation model block:
    • 17362 Now the model tree provides extensive details on model’s inputs and outputs — variable’s name, brief description, physical quantity, measurement units, variability.
    • 17387 Added the ability to export approximation models to binary formats. See Model Export section for details.
    • 17489 Added the ability to calculate the coefficient of determination (R-squared error metric).
  • Approximation model block, CAD blocks:
    • 16705 Added the ability to select multiple items in the model tree. Selected items are automatically shown on the Variables or Properties panes.
  • DRCompressor, DRDecompressor blocks:
    • 17255 Added file name validation in the Export model dialog for models exported to the Octave format.
  • CSVGenerator, CSVParser, Text blocks:
    • 16194 Unified the list of supported encodings.
  • CSVGenerator block:
    • 16159 Renamed some options in block configuration.
    • 16159 Updated the list of available settings for Line endings option.
  • CSVParser block:
    • 16159 Renamed some options in block configuration.
    • 16159 Updated all option descriptions.
    • 16159 Improved validation when configuring the block, in particular, clarified error messages shown on the Preview tab.
  • ANSYSWorkbench block:
    • 17226 Added the support for ANSYSWorkbench 19.0.
    • 16705 Improved filtering for parameter names and their descriptions in the project tree.
  • FloEFD block:
    • 17257 Added the support for FloEFD Standalone 17, FloEFD 17 for PTC Creo, and FloEFD 17 for NX.
  • Program, PythonScript blocks:
    • 16619 Added the ability to filter values of runtime environment variables in block configuration.
  • SolidWorks block:
    • 17257 Added the support for SolidWorks 2018.
  • Text block:
    • 16562 Introduced automatic syntax highlighting in templates for some input file extensions.
    • 16562 Made quick selection heuristic more user-friendly.
    • 16985 Updated icons for read and write operations when selecting values in the template.
    • 16194 Added a new Line endings option specifying the line endings to use in files output by the block.
  • GUI:
    • 17016 Added the ability to maximize all block configuration dialogs.
    • 17462 Added the ability to place the blocks in a workflow in a horizontal or vertical layout. Diagonal and Energy optimization layout modes are now deprecated and no longer accessible from the view toolbar.
    • 17253 17218 Removed the deprecated Cache, Reduce, TextGenerator, and TextParser blocks from the Block Library.
    • 17030 17062 Improved the display of long block names in link tooltips and workflows.
  • General:
    • 12994 Added a command in vector value editors which fills a vector with values taken from an arithmetic progression with specified properties.
    • 16679 Added a command line option to run a workflow without using pSeven’s graphical user interface. See Command-Line Workflow Integration for details.
    • 17332 Added a third-party Python library to run FMI models (Functional Mock-up Units). See section Open Source Components for details on FMPy and other updated components.


Compatibility Issues

This release adds an upper limit for the training sample size in the GP and HDAGP Approximation techniques: the maximum supported size is now 4000 points. If the sample size exceeds the limit, these techniques can no longer be used — for example, the ApproxBuilder block can stop with an error. If you have used the GP or HDAGP techniques on large samples, it is recommended to switch to the SGP technique.

A related change is the new upper limit for the GTApprox/SGPNumberOfBasePoints option: the maximum valid value is now 4000, while previously it could be set to any integer. In a situation when pSeven finds this option set to a value higher than 4000 — for example, if you open a workflow with such a setting created in an older version — this value is automatically reset to the new maximum. This reset ensures that existing workflows and models remain compatible with the new pSeven version, but if you re-run an older workflow or re-train an older model in pSeven 6.14, the resulting model may be slightly different.

pSeven 6.14 also introduces certain changes in approximation options, which do not create compatibility issues but may alter its behavior or results in some cases. The following options, which previously used a fixed default value, are now tuned automatically by default (use the newly added Auto value):

It means that in version 6.14 and above the actual defaults of these options may be different compared to the previous versions, since the default values are now determined automatically during training. However, in most cases this should improve model quality due to better option tuning by default.


  • Workspace:

    • 17261 Fixed wrong line endings behavior in example descriptions on Windows.
    • 17219 Fixed incorrect behavior of the Project description tab not being shown if there is no file in the current project directory or if such a file is empty.
  • Workspace, Analyze:

    17219 Fixed project-related images not being shown as changed when they are refreshed using the Refresh project description and Update report buttons.

  • Edit, Run, Analyze:

    • 17244 Fixed incorrect behavior of the tab bar when reordering tabs.
  • Run:

    • 16898 Fixed a bug which caused the inability to reassign a default value to some options being added as workflow parameters.
    • 17046 Fixed a bug due to which it was sometimes impossible to see a few last lines in the run log after running a workflow containing Program, Python script, or Echo blocks.
    • 16917 Fixed incorrect behavior of a Stop workflow button for all workflows except for the one that is running.
  • Analyze:

    • Sample viewer:
      • 17329 Fixed incorrect calculation of correlation coefficients in some cases, in particular when the sample contains constant components or duplicated points.
      • 17263 Fixed tooltips on the Correlations tab displaying incorrect (rounded or NaN) values.
    • 2D plot:
      • 16997 Fixed a bug which resulted in wrong data representation on plots with similar configuration settings.
    • Column chart, Histogram:
      • 17277 Fixed incorrect tooltips in terms of displayed values and their placement on the plots.
    • Column chart:
      • 16910 Fixed improper handling of data with negative values when switching to a log scale in viewer configuration.
    • Histogram:
      • 17277 Fixed incorrect manual axis range validation in viewer configuration dialog.
      • 17277 Various minor visual fixes.
    • Models pane:
      • 17321 Fixed the inability to export a model as an Excel document with a linked DLL, if the export directory is specified using the Export model dialog.
      • 17475 17230 Fixed the inability to work with composite models containing categorical variables.
      • 17260 Fixed the incorrect behavior of Import from file… command due to which it was sometimes possible to import a new model while the previous one was still loaded.
    • Model validator:
      • 17066 Fixed a bug due to which it was impossible to open a model with incomplete model information.
    • Model explorer:
      • 16580 Fixed incorrect placement of plot legends.
      • 16580 Fixed wrong sizes of some plots.
      • 16580 Fixed a bug due to which it was impossible to properly select a model on the Models pane if only one model was added.
      • 16666 Fixed incorrect behavior of some plots after zooming.
      • 16684 Fixed an unwanted error message being shown on the Issues pane when opening the model with no training sample.
      • 17233 Fixed model inputs and outputs being incorrectly displayed on the plots after changing the model order in viewer’s configuration.
  • Model builder, ApproxBuilder block:

    • 16174 Fixed incorrect clustering of data samples having tensor structure in the Mixture of Approximators technique, which could lead to inability to train a model if GTApprox/MoATechnique is set to TA or TGP.
    • 17450 Fixed a bug in the Mixture of Approximators technique, which could cause an unhandled error when GTApprox/MoATechnique is set to iTA.
    • 17404 Fixed a numerical issue which could cause an error when using the Gaussian Process technique with a small training sample containing a constant input and specifying different sample point weights.
    • 17353 Fixed a slowdown of the Tensor Approximation technique in the case when some tensor factors are processed with the HDA technique and GTApprox/MaxAxisRotations is used.
    • 17669 Fixed incorrect (sometimes crashing) internal validation for Tensor Approximation models in the case when the training sample contains ambiguous points.
    • 17673 Fixed a crash in internal validation when all point weights are set to 0.
    • 17436 Minor internal fixes in some of the approximation algorithms.
  • ANSYSWorkbench block:

    • 17357 Fixed a project upgrade bug due to which removing variables in block configuration could cause an error.
    • 17341 Fixed the block’s inability to build a project tree if the log contained error messages from previous runs.
  • Approximation model block:

    • 17351 Fixed a bug due to which it was impossible to run a workflow after the block’s upgrade.
    • 17412 Fixed the block being unable to output 95-th and 99-th percentile validation errors.
    • 17319 Fixed the inability to set the appropriate export format in block configuration as it returned to the previous setting.
  • Composite block:

    • 17075 Fixed incorrect rendering of HTML and CommonMark text in tooltips on the block information panel.
  • Excel block:

    • 17301 Fixed a bug which resulted in matrix variables being incorrectly transposed on writing to a disjoint range.
    • 17301 Fixed incorrect behavior of the Add variable dialog when a disjoint range is specified in an incorrect format.
    • 17416 Fixed improper behavior when writing data from a single variable to a disjoint range.
    • 17301 Various other minor bugfixes.
  • NX block:

    • 16971 Fixed the bug resulting in a failure to run the block with NX 11 on Windows as it requested for a developer’s license.
  • Program, PythonScript blocks:

    • 17306 16619 Fixed incorrect validation for the names of environment variables on the Environment tab in block configuration. Also fixed several minor inconsistencies related to duplicate names validation.
  • Text block:

    • 16152 Fixed the inability to select a variable in the template after a quick selection pop-up was once closed using Esc.
  • GUI:

    • 17388 Fixed a bug due to which some GUI buttons failed to work properly after being held for a few seconds.
    • 17388 Fixed focus indicator displacement that occurred when opening some dialogs.
    • 17448 Fixed Home, End and arrow keys not working in tables.
    • 17214 Fixed incorrect data representation in the Autolink dialog on Windows.
    • 17254 Fixed images with relative paths not being shown when editing HTML or CommonMark text in a value editor dialog.
    • 17254 Fixed StringScalar value editor sometimes not switching to the HTML mode automatically when editing a HTML-formatted value.
    • 16642 Fixed incorrect rendering of tooltips for some data values.
    • 17444 17448 Various minor fixes in usability and behavior.
  • General:

    • 17481 Fixed a possible crash when working with long parameter trees (several hundreds of nodes) in integration blocks. Very long trees are now visually divided in parts containing at most 500 nodes.
    • 17400 Fixed a rare bug due to which pSeven for Windows could not start if there is a registered MIME type with a Unicode extension in the system.
    • 15553 Properly prohibited running two or more pSeven Agent instances with different settings on the same host.
    • 16964 Fixed a number of minor issues with upgrading workflows created in previous versions of pSeven.
    • 16505 Fixed the issue with the incorrectly set working directory in the Run script dialog.

pSeven 6.13 Service Pack 1

Release date: 5 June 2018.

This release includes pSeven Core (MACROS) 6.13 Service Pack 1.

New Features

  • When training an approximation model, pSeven can now automatically apply log transformation to the training outputs, which improves model accuracy in cases when values of some outputs are exponentially distributed. This feature is available in the ApproxBuilder block and in the manual configuration of the Model builder tool in Analyze. For details, see the GTApprox/OutputTransformation option.



  • ANSYSWorkbench block:
    • Fixed a project upgrade bug in pSeven 6.13 due to which incorrect data types were set for output variables (output ports) in the ANSYSWorkbench block when opening a project created in previous versions of pSeven.
  • Approximation model block:
    • Fixed a project upgrade bug in pSeven 6.13 due to which the Approximation model block could be removed from a workflow when opening a project created in previous versions of pSeven.
  • NX block:
    • Fixed a bug in file export which could corrupt the document if it is exported to a file which already exists. The block will now correctly overwrite the existing file.
  • General:
    • Fixed a bug which allowed to open the same project in two or more instances of pSeven due to incorrect project locking.


pSeven automatically creates a backup copy of project contents on every upgrade. These backups are saved to the hidden .backup directory inside the project. To fix a project which was corrupted by the upgrade bug in pSeven 6.13, you should first restore it to the pre-upgrade state using the latest backup copy (this brings the project back to its old version). After restoring the backup, re-open the project in pSeven 6.13 Service Pack 1 and let it upgrade; do not open the restored project in pSeven 6.13.

pSeven 6.13

Release date: 16 May 2018.

This release includes pSeven Core 6.13.


This release discontinues the support for Windows Vista. It also changes a few system requirements: screen resolution 1280×1024 or greater is now recommended due to the increased font size in GUI, and 32 bit color depth is no longer required to run pSeven under Windows. See the updated System Requirements section for full details.

New Features

  • Edit:
    • Redesigned the Block library pane, added block icons and keyboard navigation support. Changed block grouping and added new tags indicating block development status — see the updated Block Library section for details.
    • You can set colors for Composite blocks in a workflow using the b_edit_setblockcolor button on the edit toolbar.
    • Tooltips for Composite blocks that have a non-empty description now show the description contents. Note that Composite block descriptions support the CommonMark syntax and can contain formatted text and images.
    • CommonMark is now also supported in port and parameter descriptions.
    • Added a white background view mode, which can be toggled with the b_edit_whitebg button on the view toolbar.
  • Run:
    • You can now open block configuration from Run by double-clicking a block (pSeven will automatically switch to Edit and open the block’s configuration dialog).
  • Analyze:
    • Added a command to copy the training sample data from an approximation model to a report database so this data can be used with other analysis tools. See section Extracting the Training Sample for details.
  • ANSYSWorkbench block:
    • Added the support for ANSYS Workbench projects with enabled parallel calculations.
    • Two or more ANSYSWorkbench blocks in a workflow can now run simultaneously if they load different projects, and your ANSYS Workbench license allows running several application instances.
  • Excel block:
    • The block can now read values of variables from disjoint selections and write data from a single variable to such a selection. For example, it allows you to set a vector variable using two or more cell ranges as data sources.
    • You can run document macros from the block’s configuration dialog, for testing purposes.
    • Added a per-variable setting for matrix variables which specifies whether the matrix should be transposed upon reading or writing.
  • FloEFD block:
    • Added the support of the Formula and Table dependency types for initial data and boundary conditions.
    • The block can now be configured to rebuild the project and run batch results processing without running calculations. See the Analysis actions option for details.
  • GUI:
    • pSeven can now be switched to the fullscreen mode: select File ‣ Toggle fullscreen from the menu or use the F11 hotkey.
  • General:
    • Core algorithmic blocks in pSeven for Windows (Optimizer, ApproxBuilder, Design of experiments, DFBuilder, DRBuilder) and the Model builder tool in Analyze now support systems with more than 64 logical processors or several processor groups (see the Processor Groups section in Microsoft Docs for details). In general, it means that pSeven 6.13 shows significant performance increase when using high degree parallelization on Windows systems with more than 64 logical processors, compared to previous versions of pSeven for Windows, where the ineffective distribution of threads among logical processors could negatively affect performance.

Updates and Changes

  • Edit:
    • The Workflow tree pane is now more functional: can be toggled on and off, resized, can be used for workflow navigation.
    • Added an editor for workflow descriptions, similar to the project description editor in Workspace.
    • Improved the layout and greatly increased the refresh speed of the block information panel. The panel now updates instantly even when selecting a block with hundreds of ports.
    • Added port name filters on the Ports tab in block configuration dialogs.
    • The Configure file dialog now selects the project file origin by default (see Run-time File Configuration).
  • Run:
    • To avoid confusion, consecutive spaces are now prohibited in names of parameter presets.
  • Analyze:
    • Model validator improvements:
      • The Test sample pane now shows names of model inputs and outputs if they are available from the model.
      • Added commands to quickly select model training or test samples as validation data sources. These commands are usable only when the training (test) data is saved to the model.
      • The model’s training sample is no longer automatically selected as the validation data source.
      • General usability improvements.
    • Plots in Model explorer are now sorted by names of model outputs from top to bottom.
    • Added the support for multiple datasets to parallel coordinates plots, along with some general improvements in their usability.
    • Improved the CART algorithm which is used to split data samples into training and test subsamples (see Split Data). The test sample points now better cover the function (response) space.
    • To prevent errors, commands in the Models pane context menu are now disabled while importing a model from file.
    • Names of data series in a report database now cannot contain leading or trailing whitespaces.
  • ANSYSWorkbench block:
    • The block now processes batches of input parameters. Due to this, its input ports that accept values of project parameters now have vector types. This change does not affect compatibility with existing workflows where scalar values are sent to the block, since any received scalar is automatically converted to a vector containing a single element.
    • If you use the Error handling behavior option, you can assign scalar values to output ports, and the block will use these values to replace results of failed calculations in its output batches. See the option description for full details.
  • CATIA, Creo, Kompas3D, NX, SolidWorks blocks:
    • Added the export format setting for temporary files.
  • CATIA, Creo, FloEFD, Kompas3D, NX, SolidWorks blocks:
    • Added a way to temporarily disable changing the value of some document parameter mapped to a port without removing the mapping or links connected to this port. Now a None value received to RealScalar port means that the block will not change the parameter value when running. Also, the listed blocks automatically assign the None value to all their RealScalar input ports. Note that if such a block is nested into a Composite block, ports of the Composite blocks created by uplinking inherit these assigned values from the nested block (see the Composite block’s changelog below). This behavior is primarily intended for the purposes of testing your simulation and evaluation blocks while designing a workflow.
  • ANSYSWorkbench, CATIA, Creo, FloEFD, Kompas3D, NX, SolidWorks blocks:
    • Added the Copy value button to the Document tree pane so values of parameters can be quickly copied from the loaded document to the clipboard.
  • Optimizer block:
    • Optimization results are now always sorted by values of variables.
  • Design of experiments block:
    • Improved performance and stability.
    • Added an option to include the initial input sample into the final generated DoE.
    • The Adaptive design technique no longer ignores the points from the initial sample that are outside of variables’ bounds.
    • Results of the Adaptive design technique now include the points with NaN response values as infeasible designs.
    • Fixed option descriptions for the Orthogonal array technique.
    • Optimized block’s configuration dialog window size.
  • Approximation model block:
    • If you use the Error handling behavior option, you can assign scalar values to output ports, and in case of failure the block will generate output data arrays filled with these values. See the option description for full details.
    • Multiple validation error types can now be selected on the Advanced tab.
    • Simplified the model export settings.
    • Added ranges of variables and responses to the model summary. Enabled copying text from the Model summary pane.
  • ApproxBuilder block:
    • Added the sample_names port to set names for variables and responses of the model trained by the block.
  • DFBuilder block:
  • Composite block:
    • When possible, types of the Composite block’s input ports and values assigned to these ports are now automatically synchronized with types and values of the nested blocks’ ports uplinked to the Composite. This synchronization is disabled for parallel ports and one-to-many connections. Also, if you assigned values to some ports manually, these values will not be changed.
  • GUI:
    • Increased the font size on GUI elements for better readability.
    • Better compatibility with systems that lack the OpenGL support. pSeven now uses a bare minimum of OpenGL features by default. A separate (optional) OpenGL mode is added, which requires OpenGL 2.1 for the Windows version and OpenGL 2.0 for the Linux version. Running pSeven in the OpenGL mode on systems that satisfy these requirements can improve performance and GUI responsiveness. The shortcuts to launch pSeven and pSeven Runner in the OpenGL mode are found in the system menu after installation.
    • The support for X11 forwarding in pSeven for Linux, which stopped in pSeven 6.8, is enabled again. This update also improves compatibility with older versions of the X server and with X server implementations for Windows. See FAQ: How to enable X11 forwarding in Windows? for details.
    • Many dialogs that contain big tables now work noticeably faster — such as Edit value… dialogs with big matrices, long option lists, and other.
    • Various other usability and layout improvements.
  • General:
    • Better file handling prevents file access errors when pSeven starts new processes such as external solvers and other tools.
    • Removed the deprecated If, Switch, and Toggle blocks from the Block Library. You can use the Condition block to create conditional nodes in the workflow (see Conditional Branching).
    • Changed the name of the pSeven backend executable from p7.exe (p7 in Linux) to p7server.exe (p7server in Linux).
    • Updated some open source components to their recent versions, see section Open Source Components for details.


Compatibility Issues

  • Due to the fix of implicit matrix transposition applied to the Excel block, existing workflows using this block may need editing in order to work properly in pSeven 6.13. In previous versions, Excel automatically transposed matrix values when it could not “fit” them into the variable’s cell range — for example, when trying to write a 5×3 matrix to a 3×5 range. In certain cases this behavior resulted in an unwanted transposition, so it was disabled, and transposition for matrix variables is now configured manually. You may need to revise variable’s settings in Excel blocks to specify which variables the block transposes on writing.


  • Edit:
    • Fixed a bug which allowed to open multiple block rename dialogs at once.
  • Run:
    • Fixed pSeven freeze after running a workflow which outputs a big matrix to a root port. Previews for such values are no longer shown on the outputs pane in Run, so this data can be processed in Analyze only.
    • Fixed a bug due to which you could not see a few lines at the bottom of the run log when using pSeven in Windows 10.
  • Analyze:
    • Model builder:
      • Fixed incorrect behavior of the Validation type hint when test data is specified to the builder or validation is disabled with the Internal validation hint.
      • Fixed incorrect SmartSelection behavior when training a model where all variables are categorical and the Exact fit requirement is added.
      • Fixed a bug due to which wrong indices of categorical variables were stored in model information.
    • Model explorer:
      • Fixed a bug which caused incorrect behavior when changing models in the Model explorer configuration.
      • Fixed several bugs related to viewing models with categorical variables.
      • Fixed a freeze when loading a model with high number of inputs or outputs.
      • Fixed an error when trying to view a model which is currently training.
    • Model validator:
      • Fixed the inability to show plots and model information when the \(R^2\) error value could not be calculated.
    • Sample viewer:
      • Fixed column selections on the Data tab disappearing after renaming data series in the viewer’s configuration or changing their order.
      • Fixed the viewer ignoring the “Display limit of scatterplot” setting when “Draw errors on scatterplot” is set to With internal validation.
      • Fixed incorrect calculation of sample statistics when the viewer’s dataset contains data series of different length.
    • Fixed incorrect behavior when trying to exit pSeven, save or close a report while training an approximation model.
    • Fixed unexpected slowdowns when importing Excel spreadsheets with a large number of columns.
    • Fixed unwanted value format conversions when exporting to Excel (strings could be converted to numeric values).
    • Fixed the inability to render a 2D plot when [index] is selected as the data source for the y-axis.
    • The Delete command in the Models pane context menu is no longer available for models that are currently training or waiting.
    • Fixed a bug which did not allow opening a report created in pSeven 4.0 or earlier versions if the report contains a parallel coordinates plot.
  • CATIA, Creo, FloEFD, Kompas3D, NX, SolidWorks blocks:
    • Fixed a rare error while building the document tree which resulted in a “Cannot build document tree: ‘children’” error message.
  • Excel block:
    • Fixed the block implicitly transposing matrix values, for example when writing an \(n \times m\) matrix to a \(m \times n\) cell range. Transposition is now an explicit option specified for each variable by user.
    • Fixed incorrect sheet name validation in the block’s configuration dialog.
    • Fixed the Clear table command not clearing the list of name in the sheet selector.
    • Fixed the inability to select the same single cell range as a data source for two or more output variables.
  • Integration blocks:
    • Fixed potential problems when running on Windows systems where values of environment variables contain non-ASCII characters.
  • Optimizer block:
    • Fixed a bug which in some cases lead to exceeding the evaluation budget set by GTOpt/MaximumIterations.
    • Fixed incorrect calculation of budget for expensive evaluations in some cases: the evaluations of initial guesses expended the budget set by GTOpt/MaximumExpensiveIterations despite they should not.
    • Fixed incorrect caching of expensive response values when working in batch mode: Optimizer lost useful information because it did not cache the values of those responses which it did not request to evaluate, even if these responses were received.
    • Fixed a bug in handling NaN response values in batch mode which could negatively affect the solution quality.
    • Fixed an error after deleting a variable, objective or constraint used in variable grouping.
    • Fixed a bug in compatibility with older pSeven versions which caused inability to switch tabs in the block’s configuration dialog after a workflow upgrade.
    • Fixed outdated option descriptions.
  • Design of experiments block:
    • Fixed incorrect handling of NaN response values when the Adaptive design technique is used.
  • Approximation model block:
    • Fixed incorrect handling of a relative path to the model file.
    • Fixed loading of old approximation models that do not contain information about training options.
  • ApproxBuilder block:
    • Fixed a bug due to which wrong indices of categorical variables were stored in model information.
  • Program block:
    • Fixed the inability to soft stop the block when it keeps retrying a remote connection.
    • Fixed a bug due to which the block incorrectly reported remote command execution errors.
    • Fixed several option descriptions and labels in the configuration dialog.
    • Name filters on the Files and Substitutions panes are no longer case-sensitive.
  • DFBuilder block:
    • Fixed a bug in the low-fidelity model bias compensation algorithm (see GTDF/UnbiasLowFidelityModel) due to which it worked incorrectly when parallelization is enabled.
  • PythonScript and other blocks with user-defined variables: fixed a bug due to which the block could show an error about a nonexistent conflict in names of variables, when variables are added and renamed afterwards.
  • Approximation model, Const, Design of experiments, Program blocks:
    • Fixed a bug that in rare cases could crash pSeven when changing names of constants or variables in block configuration.
  • pSeven Runner:
    • Fixed several minor inconsistencies in the user interface.
  • General:
    • Fixed incorrect installation under Linux when the installation path contains spaces or special characters.
    • Fixed the inability to properly exit pSeven after the connection to license server was lost.
    • Fixed incorrect path separators in the path to pSeven which is added to system PATH when installing pSeven for Windows.
    • Fixed incorrect behavior of column sorting functions in some tables in GUI. Also, sorting by name is now case-insensitive.
    • Various minor usability fixes.

pSeven 6.12 Service Pack 2

Release date: 6 March 2018.

This release includes pSeven Core (MACROS) 6.12 Service Pack 2.

Updates and Changes

  • Edit: improved responsiveness of the user interface when opening and closing block configuration dialogs, switching tabs in block configuration, and for other common editing actions. In particular, significantly increased responsiveness in the following cases which previously caused unexpected slowdowns:
    • mass adding ports to a block,
    • configuring a “heavy” block with a large number of ports,
    • configuring a Composite block that contains some “heavy” block,
    • clicking Apply in a “heavy” block configuration dialog or closing it,
    • creating many links at once,
    • grouping many blocks to a Composite,
    • opening, editing, validating or running a workflow containing “heavy” blocks, and other similar cases.
  • Analyze: updated Mutual Information and Distance Correlation algorithms used by the Sample Viewer. Mutual Information is now calculated significantly faster. Distance Correlation calculation runs faster and has reduced memory consumption.
  • Approximation model, ANSYSWorkbench, Creo and other integration blocks: updated the components responsible for building and showing the document/model tree, greatly improving its responsiveness and usability. In particular:
    • Large trees are now built much faster.
    • Configuration dialog is no longer blocked while building the tree.
    • Keyboard navigation is enabled in the tree.
    • Name filter can now search for any part of the parameter name, and other usability improvements.



  • Workspace: fixed the inability to open a project created in previous versions of pSeven in case when this project contains a database backup. This situation is possible, for example, if previously the pSeven process was interrupted while running a workflow from this project.
  • Edit: fixed a port naming bug in blocks that allow you to add ports by mapping parameters from a document or model tree (for example, Approximation model or integration blocks). The bug made it possible to create ports with names containing forbidden characters, thus breaking block configuration. Now the forbidden characters (<>:"/\|?*) are always removed from port names. If the removal creates two or more ports with the same name, the name collision is resolved by appending numbered suffixes: (1), (2) and so on.
  • Run: fixed a bug in the Linux version of pSeven which caused inability to run a workflow that contains a large number of blocks, or some Composite blocks with high parallelization ratio and several nested blocks.
  • ANSYSWorkbench block: fixed Unicode support in system and parameter names, file paths and other settings.
  • Creo block: fixed incorrect error message in case when PRO_COMM_MSG_EXE or the Creo process communication executable option are not set properly.
  • FloEFD: fixed a compatibility bug which was discovered due to internal changes in the FloEFD API introduced by FloEFD 16.2. Due to this bug, the block was unable to get result values, so all results were set to 0.0. This bug did not affect compatibility with FloEFD versions prior to 16.2.
  • ANSYSWorkbench, FloEFD, Program (under Windows): fixed incorrect behavior of applications started by these blocks, which was caused by loading incorrect DLLs. This bug was introduced in pSeven 6.12 Service Pack 1 when fixing the crash caused by incompatible versions of Python packages (PyWin32) installed in the system; earlier pSeven versions do not contain this bug.
  • Program block: fixed incorrect SSH connection behavior in the case when some shell initialization script on the remote host contains a command which prints to stdout. The block will now correctly report this error when you test SSH connection or run the workflow.
  • Optimizer block:
    • Fixed an inconsistency in batch mode options handling, due to which even with batch mode disabled Optimizer internally formed batches of points to evaluate. In particular, this often made the block unresponsive to the manual stop command.
    • If batch mode is enabled and multiple initial guesses are sent to the block (see Re-using Problem Data), these points will also be evaluated in batches. Previously Optimizer effectively ignored the batch mode when evaluating initial points: it formed batches always containing a single point.
    • Fixed a bug due to which some of the evaluation results for points from the initial guesses sample (received to the designs port) were sometimes ignored by the solver.
    • Fixed small discrepancy between the input values of multiple initial guesses (received to the designs port) and their values sent to evaluation (output variable ports).
    • Fixed a rare memory violation when a large number of parallel threads is used by the solver.
  • Approximation model block:
    • Fixed the inability to view some of the RSM model details in the model tree (Model Details — Regression Model) and the error on running the workflow if corresponding ports were added to the block.
    • Fixed a bug which could crash pSeven when manually specifying tree nodes on the Properties pane (Advanced tab).
    • Fixed incorrect behavior of the Add properties dialog when the value specified in the “Tree node” node field cannot be converted to a StringVector.
    • Fixed a bug in Gaussian Processes model evaluation due to which model response for a single input point could be slightly different from the response for the same point passed as a part of an input sample.
  • GUI: fixed a bug which sometimes caused spontaneous resizing of GUI panels.
  • General: when you input a single string literal as a value for some StringVector or StringMatrix variable, the value is now correctly converted to a StringVector or StringMatrix with a single element.

pSeven 6.12 Service Pack 1

Release date: 18 December 2017.

This release includes pSeven Core (MACROS) 6.12 Service Pack 1.

Updates and Changes

  • Run:
    • Improved the layout of Input and Output panes which now always show the tree of input/output ports after updating the workflow configuration or after its run.
  • Analyze:
    • Improved automatic generation of data series names: the name of the port being the source of data is preferably used in data series naming.
    • Increased the size of the Report database pane.
  • CSVGenerator, CSVParser blocks: set a comma (,) as default delimiter.
  • Program block: changed the color of the error icon in the test connection dialogs.
  • General: added qtdiag (a useful tool for diagnostics) to the pSeven distribution.


  • Edit:
    • Fixed the bug which resulted in the inability to update port status on the block information panel and in tooltips after the connection to the port has been removed.
  • Analyze:
    • Fixed the bug which resulted in pSeven crash when evaluating the model with identical inputs or outputs in the Make predictions dialog.
    • Fixed the bug which resulted in data series being imported from CSV in the alphabetic order and not in the original one.
    • Fixed wrong representation of data via the ‘Format’ option in Page viewer, Parallel coordinates and Sample viewer.
  • Approximation model block:
    • Fixed the bug due to which the port’s type was not properly updated after changing the properties’ type.
    • Fixed the bug due to which the block could fail to output values to Responses.* port, if the port was not connected.
    • Fixed the bug due to which the block could fail to output Accuracy, if at least one Gradient accuracy port was not connected.
  • Optimizer block: fixed the bug which resulted in the inability to reset the initial guess value in some cases.
  • Design of experiments block:
    • Fixed the bug which resulted in the inability to run workflow with Optimal design/Pure quadratic Technique/Model settings.
    • Fixed rare misbehavior of adaptive design algorithm when it returned less points than requested.
    • Fixed the bug in the adaptive design algorithm which resulted in unexpected grouping of points in the design space in some rare cases.
  • General:
    • Fixed a bug in pSeven for Windows due to which it could crash on start if Python for Windows Extensions (PyWin32) are installed in the system.
    • Fixed the bug due to which blocks were unable to clear the sandbox under Windows if the sandbox prototype contained a file with the read-only attribute.
  • GUI:
    • Fixed string data editor sometimes not responding to the mouse scroll wheel.

pSeven 6.12

Release date: 17 November 2017.

This release includes pSeven Core (MACROS) 6.12.

New Features

  • Design of experiments block: a new Design of experiments block has been released. This block features the following updates:

    • New Adaptive design technique: generation of designs taking into account initial sample and constraints.
    • New Responses pane:
      • Each response can be one of four types: Evaluation, Constraint, Exploration and Contour. The last two types are specific for the Adaptive design.
      • It’s now possible to specify which type of function the response belongs to: Generic, Cheap, Linear or Quadratic. Adaptive design takes this information into account during generation. Each function type can be evaluated with a separate blackbox.
    • New Unfiltered designs ports output both feasible and infeasible designs, while Designs ports output only feasible ones.
    • Support of the initial sample with values of responses. Values of responses can be incomplete, that is, the initial sample can contain missing values.
    • Run options:
      • Log level option was renamed to Verbose output and is now Boolean.
      • Renamed Max parallel option to CPU cores limit.
      • Replaced Batch size option with two new options: Batch mode and Maximum batch size.
    • Changed naming rule for block ports and added the Blackbox prefix to all ports required to connect blocks which evaluate responses.

    Design of experiments replaces the old DoE block. Note that workflows containing the old DoE block or the beta version of the Design of experiments block are not upgraded automatically.

  • Approximation model block: added a new Approximation model block that allows using models built by the ApproxBuilder block or the Model builder tool (see Model Training). Main block features are:

    • When a model built with Model builder in Analyze is loaded, the block automatically creates ports for each model input and output.
    • Variables and Responses panes: manual configuration of model variables (inputs) and responses (outputs).
    • Model summary and Training options panes: the block shows useful information about the model when it is loaded.
    • Model tree pane: any piece of model information can be written to a separate output port.
    • Runtime model export is available as part of Model tree.

    Approximation model replaces the old ApproxPlayer block. Note that workflows containing the old ApproxPlayer block are not upgraded automatically.

  • Optimizer block:

    • Added the GTOpt/ResponsesScalability option that changes the computational model used by surrogate-based and robust optimization methods.
    • Added the capability to run in deterministic mode which makes the optimization process reproducible exactly. See the related GTOpt/Deterministic and GTOpt/Seed options.
  • ApproxBuilder block: added a new GTApprox/MaxAxisRotations option that enables using model gradients during training in order to obtain smoother models.

Updates and Changes

  • Edit:
    • Added the ability to sort workflow inputs and outputs by alias.
    • Moved the Workflow description button to the right toolbar as a toggle button.
    • Added ‘Run workflow’ button on the Edit screen.
  • Run:
    • Enabled monitoring for default run parameters.
    • Enabled user selection of sources in the run log to be stored between runs of the same workflow.
    • Improved pop-up notifications depending on the workflow finish status.
  • Analyze:
    • Made the background color of the plots white for a better look of screenshots in documents.
    • Added KaTeX formulas support to the Text viewer.
    • Faster model training in the Model builder.
    • Significantly reduced model reloading time in the Model explorer.
    • Reduced memory consumption for complex 3D plots/surfaces.
    • Enabled scrolling for sliders to improve the usability of the Model explorer.
    • Enabled a new error message to be shown on the Issues pane of the Model builder when it is impossible to use internal validation.
    • Improved error messaging in Import data from CSV file dialog for cases when the encoding specified in the dialog settings is other than the one of the CSV file. Extended the available encoding list in Export/Import from CSV file dialogs.
    • Improved wildcard characters support when exporting data.
    • Updated several icons in Analyze and Excel block configuration dialog.
    • Made ‘Smooth model’ and ‘Exact fit’ hints in the Model builder to be represented as incompatible to reflect similar changes introduced in pSeven Core.
    • Removed ‘Empty viewer’ validation warnings for Sample viewer and Page viewer for consistency purposes.
    • Improved the behavior of the ‘Correlations only’ flag for X vs Y plots; now the flag is inactive, if the Sample viewer window becomes too small.
    • Fixed export model to executable: .exe extension is not added to the Linux executable.
    • Modified the names and order of variables used when exporting statistics to a CSV file.
    • Disabled ‘Discard values’ and ‘Split data’ options in the menu for empty report database.
    • Enabled using the Project database pane menu for the empty project database. Made the Reduce database size command always available (except for cases when the workflow is running).
    • Improved usability of the Model details dialog when viewing models with a large number of inputs or outputs.
    • Improved representation of extrapolated points being outside the training data boundaries in the Model explorer.
    • Improved Make predictions dialog; now it shows actual input/output names taken from the model description instead of generic names like “X0”, “F0”.
    • Added the ability to filter data in the Sample viewer right from the data table without opening the Configuration window.
    • Added the ability to partially set output noise variance in the Model builder. Also the option’s value can be set as a number.
    • Added support of zero-dimensional matrices during queries recalculate operation.
    • Added the ability to set prefixes for generated data series in Make predictions dialog. Removed ‘Make prediction’ button from the Model validator.
    • Added tooltips for each data cell in the Sample viewer.
    • Enabled Enter hotkey in Run workflow with errors dialog.
    • Added the ‘Format’ option to dataset configuration in Parallel coordinates.
    • Improved the layout of the text on logarithmic axes in Parallel coordinates.
    • Removed file extension restrictions (.txt, .csv) for import from CSV operations.
    • Changed the behavior of Import data from file dialog. The improvements are as follows:
      • Unnecessary refreshes removed;
      • Only visible tab is updated;
      • Auto refresh disabled for the current tab, if the time to wait for response was more than 10 seconds;
      • Increased performance for big files on the Preview and Results tabs in import dialogs;
      • Introduced an info pop-up about auto refresh disabled.
  • SolidWorks block: added a new error message for CAD incomplete or corrupted installation.
  • Program block:
    • Introduced slight improvements in Slurm jobs handling.
    • Enhanced the informational value of the block’s error log.
  • Composite block:
    • Increased maximum ParallelizationRatio value for Composite blocks from 32 to 64.
    • Enabled ‘All files’ filter in the file dialog for the cache of the Composite block.
    • Extended the group blocks function to grouping of selected annotations and moving them inside a new Composite block.
    • Added an instance number to every log message received from blocks nested in a parallel Composite.
    • Enabled monitoring by default for new ports in the Composite block.
    • Improved deadlock detection (connecting two links to one output port should always result in execution error).
  • Optimizer block: added objective maximization.
  • Optimizer, ApproxBuilder, DFBuilder, DRBuilder blocks: changed the default behavior of GTOpt/MaxParallel, GTApprox/MaxParallel, GTDF/MaxParallel, and GTDR/MaxParallel options to limit parallelization on hyper-threading CPUs. See option descriptions for details.
  • Excel block:
    • Improved handling of Excel errors;
    • Added the ‘Clear table’ button;
    • Added the ability to work with data which size is unknown in design-time;
    • Significantly improved the overall block’s performance.
  • CSVGenerator block: added the ability to choose row delimiter between Auto and CRLF and LF.
  • Text block: added custom decimal separator support. Added the possibility to set a decimal separator for each operation.
  • GUI: added tooltips on the main application toolbar.
  • General:
    • Made autolink heuristic more user-friendly.
    • Introduced better autoalias generation for ports of some blocks.
    • Improved network paths support under Windows. Now the users can save and open projects, workflows, reports etc. to UNC paths.
    • Improved the Rename file dialog; now the file’s name is selected without the extension which is more convenient for renaming purposes.
    • Improved the algorithm of sorting the files by their names in dialogs and in Workspace.
    • Added trailing .0 for REAL values rendering.
    • Enabled ‘Maximize’ button in the Text viewer.
    • Added a new checkbox column to the Uplink dialog to facilitate uplinking the ports.
    • Added a 3rd-party pyNastran library.
    • Enhanced the stability on slow machines.
    • Updated MathGL to the latest version (2.4.1).
    • Improved a license selection dialog in the installer.
    • Improved error messages for worklow/report documents created using the older/newer version of pSeven.
    • Introduced some annotation editor improvements. Now the editors: have borders, are semi-transparent like annotations themselves, work properly, they also underwent a scale bugfix and layout improvement.
  • Agent: added the ability to install pSeven Agent under Windows for users with no administrator rights.


  • Added a new example 2.18 General - COMSOL Integration showing how to solve the task of current optimization in superconducting magnets.
  • Added detailed Program documentation. Updated option descriptions. Hided options ‘Interpreter’ and ‘Status check interval’ from the block’s GUI.
  • Updated the example 2.09 General - MATLAB Integration to change the logic for the working directory and search path handling.
  • Updated example projects and workflows where the deprecated DoE block was used.
  • Updated screenshots for the Composite, Optimizer, Condition, UQ and Distribution blocks.
  • Removed the deprecated Toggle block documentation.
  • Added a new example for 2.19 General - MathCAD 15 Integration.
  • Updated section Known Issues.
  • Removed section References.
  • Updated section pSeven Data Types (moved to Typesystem).
  • Moved section Viewer and Editor Hotkeys to the top level in the table of contents.
  • Updated section Block Configuration: Deterministic option was described (common for 7 blocks).
  • Updated section describing Model builder modes (see Model Training). Added information about SmartSelection (description, hints).


  • Edit, Run:
    • Fixed the bug which resulted in a wrong workflow representation after some old workflows have been upgraded in pSeven 6.11.
  • Run:
    • Fixed incorrect behavior of the running workflow being blocked for editing with a delay, which may result in synchronization errors.
    • Fixed the bug which resulted in the immediate stop of a workflow and an error message ‘Cannot unlock mutex - it is not locked’ being displayed; the bug affected FloEFD and probably some other CAD/CAE blocks.
    • Fixed a rare crash which occurred when the running workflow was terminated by the user.
    • Fixed incorrect validation of option values being added as workflow parameters which resulted in the inability to assign a default value in the Parameters pane.
    • Fixed the bug due to which it was impossible to run more than 60 blocks under Windows.
  • Analyze:
    • Fixed the bug which resulted in a wrong source type selection for the training sample in the Model validator.
    • Fixed the bug which could cause a freeze when opening string-type data series containing Unicode symbols in Sample viewer or Page viewer.
    • Fixed GUI errors that occurred when closing the project.
    • Fixed an error that could occur when scrolling very small tables in the UI.
    • Fixed a rare crash that could occur when multiple 3D plots were redrawn simultaneously.
    • Fixed the bug which resulted in the incorrect behavior of display limit option for scatter plots in case of different X vs X and X vs Y plot dimensions.
    • Fixed the bug which resulted in pSeven freeze after interaction with the console when training a model.
    • Fixed the bug due to which Manual axis range option was ignored, if only upper or lower bound was specified.
    • Fixed the bug due to which column selection may be lost when opening the Configuration dialog.
    • Fixed the problem with the Retrain button in the models pane of the Model builder that incorrectly restored options values from the original model.
    • Fixed inconsistent behavior of the Rename command for models and data series.
    • Fixed the bug with index misrepresentation in the Page viewer when switching from a dataset with more than one data series to a dataset with only one data series.
    • Fixed the bug which resulted in a focus indicator failure when scrolling to the end of the Sample viewer table.
    • Fixed a wrong multi-selection (Shift + Arrow up) behavior in the report database.
    • Fixed the bug due to which filtration settings were ignored in Histogram and Column chart.
    • Fixed the improper behavior of the inverted range filter in the Column chart.
    • Fixed a rare core dump during 3D plots refresh.
    • Fixed the bug which resulted in the Model explorer failure to show the model graph with NaN values in AE.
    • Fixed the bug which resulted in pSeven crash when using the Split data… command for duplicates.
  • Block configuration:
    • Fixed the bug due to which it was impossible to set up an alias for the input monitored port.
    • Fixed a bug due to which the Sandbox was not properly cleared, if workflow execution failed.
  • Blocks: fixed a rare error that occurred when switching between block configuration and ports/options settings tabs.
  • Text block:
    • Fixed the bug due to which the block may produce the incorrect UTF-16 file.
    • Fixed an error that occurred when saving the script with multiple lines statement.
  • Program block:
    • Fixed a broken ‘Time limit’ option on the Cluster settings pane.
    • Fixed the problem with LD_PRELOAD value set by pSeven GUI that affected the actual environment of the block.
    • Fixed the problem with running a script or a command using pSeven Agent under Windows so that nothing is to be removed from PATH environment variable.
    • Fixed the bug due to which active preset values affected edited block’s GUI.
    • Fixed the bug which resulted in the inability to run the block remotely via SSH, if the Remote sandbox was set to “.”
  • PythonScript block: fixed a rare pSeven crash that occurred when changing the block’s port settings with a default value being specified.
  • Creo block:
    • Fixed the bug due to which it was impossible to build a document tree, if the assembly and one of its parts had the same names.
    • Restored Creo 2.0 support.
  • Excel block: fixed a rare bug which resulted in the block’s improper execution when many parallel instances of the block were running at the same time.
  • Composite block: restored monitoring of externally unconnected ports of a parallel Composite block.
  • General:
    • Restored KaTeX syntax support in workflow annotations.
    • Fixed the bug which resulted in pSeven being unable to start under Windows when PYTHONPATH environment variable was modified by external program in Windows registry.
    • Fixed a problem with pSeven being unable to run the local license server on a Windows machine without MSVC runtime.
    • Fixed the bug due to which headers were not shown in some infotips.
    • Fixed a rare bug which resulted in the invalid value being contained in presets.
    • Fixed the bug in Workflow configuration dialog due to which it was impossible to properly set aliases and descriptions, if the workflow contained blocks with the same name, but in different Composite blocks.
    • Fixed the Dict data editor not performing checks for unique key names which may result in a loss of data entered by the user.
    • Fixed the bug due to which it was impossible to set an alias to the block’s port, if the block was nested in the Composite and was copied (and pasted) with the parent Composite block.
    • Fixed the bug with manual location select in the root of the Windows disk.
    • Fixed the bug which disabled scrolling in some edit dialogs after the dialog window has been moved.
    • Fixed the bug due to which the block’s link icon sometimes remained intact after the block was deleted.
    • Fixed the Filesystem pane being not updated every time when a project is created.
    • Fixed improper behavior of directory selection in file dialogs and panels. Fixed the bug due to which the selected directory or file can change after pressing the Refresh button.

pSeven 6.11

Release date: 1 August 2017.

This release includes pSeven Core (MACROS) 6.11.

New Features

  • Workspace:
    • Added the ability of opening pSeven documents located outside the current project. When you try to open a workflow or report that is not a part of the currently open project or is incorrectly placed in a project subdirectory, the Open workflow dialog now offers to resolve this issue.
  • Edit:
    • Introduced new icons for blocks.
    • Added the ability to edit workflows when calculations are in progress (except for the one which is running).
  • Analyze:
    • Introduced a new correlation matrix with an opportunity of input and output selection.
    • Major improvements of data filters in various viewers:
      • added the ability to filter string dimensions;
      • added batch operations for set filters (dataseries), ranges (axes), selections (axes);
      • added the ability to exclude NaN and empty values, as well as unwanted data series from results.
    • Enabled a feature to deal with samples of different size for Sample viewer.
    • Added the ability to copy data from Sample viewer to the clipboard using the Ctrl C hotkey.
    • Added an option to set the spacing between axes in the parallel coordinates plot.
  • Added the Design of Experiments block (beta). It generates data samples with specific properties and can evaluate responses on generated designs. The block is currently under development.
  • Optimizer block: added Direct Surrogate-Based Optimization (Direct SBO) technique which uses an alternative criterion to select new points for evaluation (see section Direct SBO for details). The new technique can be useful for constrained or multi-objective optimization problems when searching for non-trivial trade-offs between different expensive objectives and constraints.
  • Excel block:
    • Added ‘Max steps’ option. When its value is set to 1, it allows the block to work with an Excel document updated by external applications.
  • Predictive modeling tools:
    • Added the ability to assign unique names to DLLs to avoid accidental collisions when running multiple models.
    • Added support for both the 32-bit and 64-bit versions of Excel.
    • Added the ability for RSM models to be exported as formulas for Excel.

Updates and Changes

  • Workspace:
    • Added the ability to select the particular screen (Edit or Run) to be opened when clicking on the workflow link in the Project description.
  • Edit: disabled Cut and Duplicate commands for running workflows.
  • Run:
    • Improved pSeven runtime performance when there are too many blocks in the workflow.
    • Enabled double-click navigation for the Statistics tab.
    • Added a warning message about simultaneous use of the same sandbox by parallel instances of the block during parallel Composite run to be shown on the Issues pane.
    • Enabled a tree view for parameters on the Inputs and Outputs panes.
  • Analyze:
    • Improved validation in Import data from CSV file dialog.
    • Massive improvements in data import from Excel.
    • Improved CSV import for big files.
    • Added support for multiline headers.
    • Accelerated import of big files.
    • Reduced the min column width of the descriptive statistics table in Sample viewer.
    • Added the index column in the Report database.
    • Enabled a quick search for Models and Data series tables.
    • Enabled tooltips for 2D plots showing all dimension values.
    • Enabled batch operations for Remove dimension command in the Data settings table of Model builder.
    • Improved performance of Model validator and Model explorer to make them show selected models and not the data series.
    • Added value tooltips for points, if markers are disabled in plot configuration.
    • Minor GUI improvements in Model builder.
    • Added a divider to separate Duplicate and Remove menu items.
  • Block configuration: enabled hierarchical ports name support for the Ports tab and information panel.
  • CAD blocks: changed the recommended way of selecting the document that references other files.
  • ApproxBuilder block: improved error messaging when the output file path doesn’t exist.
  • ANSYS Workbench block:
    • Added the ability of running user scripts before or after the simulation.
    • Added the Clear Generated Data option.
    • Added support of ANSYS Workbench 17.1 and 18.
  • NX block:
    • Now the block launches Siemens NX in a separate process which enables more stable work and better error handling.
    • Added new option ‘NX operation timeout’ to set maximum time during which the block waits for NX process.
    • Enabled Manual stop for NX block.
    • Added new option ‘Trace NX API calls’ which facilitates troubleshooting.
    • Option ‘Path to DLLs’ was renamed to ‘Path to libraries’ and now works both on Windows and Linux.
    • Selecting a temporary file with initial data as the document file is now forbidden. If you use an assembly which references other files, pack them manually into zip archive. The block doesn’t need a file with information about main assembly in archive any longer.
    • Added support of Siemens NX 11.
    • Document tree build now takes less time for huge assemblies.
  • PythonScript block: changed policy for the block’s current directory which should be specified before each run, if the Keep globals option value is False.
  • Program block:
    • Improved handling of relative paths for files of project or sandbox origin.
    • Minor layout improvements.
  • Creo block:
    • Improved error handling behavior to enhance stability and responsiveness.
    • Introduced more detailed and clear error messages.
    • Ensured correct Creo behavior when several versions are installed.
    • Increased speed of document tree build and changed the structure of the document tree.
    • Added the ability to export any model from assembly.
    • Added the ability to export model views as images.
    • Added support of Creo 4.0.
  • FloEFD block:
    • Added two new options ‘FloEFD.Product’ and ‘FloEFD.Version’ instead of old ‘Version’ option. They allow switching between supported products (FloEFD Standalone, FloEFD for PTC Creo, FloEFD for NX) and different versions and automatically select the appropriate product according to the selected document file.
    • Switched off Automatic rebuild option from API.
    • Improved FloEFD performance for cases with exceeded timeout and when ‘Max steps’ is reached.
    • Added support of FloEFD for PTC Creo and FloEFD for NX.
    • Added support of non-constant parameters, initial conditions, xlsx files.
    • Usability improvements in the document tree.
    • Introduced new more informative port names instead of default ones.
    • Added support of host, number of cores and other solver options.
  • Excel block:
    • Improved performance due to suppressing ‘Do you want to save changes’ message.
    • Added an ability to read and write None and NaN values.
  • Optimizer block: added the ability to pass designs with partially calculated responses to the block.
  • SolidWorks block: added the image export feature.
  • Formula (beta) block: implemented a declarative computation model where the sequence of expressions in a block is not crucial for calculation results. In this case, circular dependencies between expressions are forbidden, but it is allowed to use formulas of the type \(y=F(y)\), where F is a certain function, e.g. \(y=-y\).
  • GUI:
    • Block and workflow layout restyling. Updated all examples due to the latest changes in pSeven visual style.
    • Minor visual and usability improvements in various dialogs, viewers and editors.
  • General:
    • Enabled HTML-formatted text in port descriptions.
    • Increased the default font size of project and workflow descriptions.
    • Added the ability to insert a file path into the StringScalar value editor.
    • Added pSeven license information to the About window.
    • Implemented a busy indicator for the tab of running workflow and model training.
    • Changed the logic for deleting temporary files created during the workflow run: now such files will be removed irrespective of the workflow finish status.
    • Implemented the search and replace function with hotkeys in the Project description editor, Run script dialog, PythonScript, Text and Program blocks.
    • Enabled warning messages to be shown on the Issues pane, if there are any unconnected output ports.
    • Converted formulas in example descriptions from images to \(\TeX\).
    • Limited internal window maximum size to the current size of pSeven window.
    • Added an ability to use hyperlinks to refer to files inside the current project directory.


  • Updated the 3.6 Three-Section Beam example for compatibility with Nastran 2013.
  • Updated section License Setup. Added a note on the necessity to update the vendor daemon and instructions on how to update.
  • Updated Tutorials. Removed outdated screenshots and amended some sections to make them more relevant.
  • Added a new section Known Issues. It details the known issues in pSeven along with possible solutions or workarounds (if any).


  • Workspace:
    • Fixed an error that could occur when closing several ‘Upgrade workflow’ message boxes by pressing ESC quickly numerous times.
    • Fixed an error that could occur when opening a new project and quickly pressing ESC at the same time.
  • Edit:
    • Fixed block startup counters being reset to 0 when clicking on the annotations icon on the view toolbar.
    • Fixed the incorrect behavior of Fit to size command for zoomed workflow.
    • Fixed incorrect export of the Composite block with enabled parallel execution.
  • Edit, Run:
    • Fixed an error resulting in Dict values conversion to string ones.
  • Run:
    • Fixed the inability to change a parameter value in Run, if this parameter is an alias for two or more different block options.
    • Fixed a bug due to which it was impossible to change the file selected once in the Parameters pane in Run.
    • Fixed a bug resulting in ‘Wait timeout exceeded’ error message being mistakenly generated by some integration blocks.
    • Fixed block counters being reset to 0 during and after interaction with the workflow.
    • Fixed a Windows-specific bug which could cause a freeze after running the Program block of the Script type with incorrect SSH settings.
    • Fixed incorrect behavior of a ‘Finish block step’ button not being displayed during the workflow run when hovering over a block.
  • Analyze:
    • Fixed the problem with incorrect data type conversion when running Create new data series… command.
    • Fixed filter validation and some minor bugs in Model builder.
    • Fixed an error due to which the invalid ‘Outputs’ variable was displayed in the dependency report.
    • Fixed incorrect rendering of colors in 2D plots expanded from the Correlation tab.
    • Fixed a bug due to which it was impossible to retrain a model in the manual mode in some rare cases.
    • Fixed an issue with redundant settings of fixed-width columns when importing data from a CSV file.
    • Fixed incorrect integer validation in Sample viewer.
    • Fixed a bug which could cause an unexpected freeze when importing data from a CSV file.
    • Fixed incorrect behavior of Create new data series… command in parallel coordinates plot.
    • Fixed incorrect data layout on the Dependency tab in Sample viewer.
    • Fixed pSeven crash when adding data to report.
    • Fixed pSeven crash when reducing a huge database size while the application is running.
    • Fixed an error that occurred when trying to remove the same data from the project database twice.
    • Fixed a bug in CSV import settings.
    • Fixed incorrect validation of Model validator configuration.
    • Fixed incorrect dataset ordering in Model builder.
    • Fixed option validation in Model builder.
    • Fixed errors with data series import.
    • Fixed cross settings validation in parallel coordinate plots.
    • Fixed a bug due to which newly opened report tabs did not show tooltips.
    • Minor bugfixes in Sample viewer.
    • Various other minor bugfixes in viewers and tables.
    • Fixed a bug due to which it was impossible to export data to file in rare cases.
  • Block configuration: fixed an error due to which the paths specified either in the Sandbox or Prototype field were removed when switching between the tabs in configuration dialogues of some blocks.
  • Distribution block: fixed the inability to remove variables properly after reopening.
  • Creo block:
    • Fixed incorrect behavior of the Open document command that used to open only the latest version of the document.
    • Fixed the problem with export to .prt format, when an empty temporary directory remained after each export.
    • Fixed incorrect behavior when selecting a document file without an extension.
    • Fixed a bug when the block failed to connect to already opened Creo application.
    • Fixed incorrect behavior of Creo that was unable to work properly after workflow termination.
    • Fixed incorrect behavior of the workflow that failed to open Creo application again and proceed with it after it was manually closed.
    • Fixed incorrect behavior of the block which could connect to Creo application run by another user, if it had proper rights.
    • Fixed a bug due to which the parameters subtree might be empty, if some model in assembly didn’t have parameters.
    • Fixed a bug which resulted in a failure to open the document with ‘Invalid number of parameters’ error when both Creo 2.0 and Creo 3.0 were installed and used with the block.
  • Excel block: Fixed a bug when Read variable and Write variable commands generated block variables with wrong “Cells” property, if cells were selected from columns which names contained more than one letter.
  • CSVGenerator block: fixed CSVGenerator being unable to create a new file, if the input port file was marked as parameter.
  • CSVParser block: fixed a synchronization error which could occurr when more than one character was entered as escape character.
  • FloEFD block:
    • Fixed incorrect behavior for cases with mapped RealVector node when the output port was created without any value.
    • Fixed an error that occurred when building the document tree with different goals in project and project result.
  • Kompas3D block: fixed an issue due to which in rare cases workflow with Kompas3D block could not finish.
  • NX block: fixed a bug due to which pSeven hang on exit after the document tree was built in the block.
  • Program block:
    • Fixed the invalid encoding of error messages under Windows.
    • Fixed incorrect behavior of ‘Time limit’ parameter for cluster jobs.
    • Fixed a bug in block’s configuration that occurred when testing a SSH connection with empty host.
  • Program block, Constant block: fixed a bug due to which it was impossible to assign Undefined value to the port instead of the default one.
  • UQ block: fixed a bug in block configuration due to which it was impossible to set distribution parameters.
  • UQ, Program blocks: fixed a bug in workflow compatibility resulting in UQ and Program blocks being removed when opening workflows created in versions 6.9 and below in the current version of pSeven.
  • Program block:
    • Fixed the terminate workflow function not working sometimes, if the Program block runs child processes.
    • Fixed the problem with simultaneous adding/removing an item with the same name in two tables.
  • PythonScript block: fixed a bug that caused a crash when changing the output port settings.
  • Optimizer block:
    • Fixed a bug which could cause accidental removal of vector elements.
    • Fixed a bug when the block tried to write an empty matrix to the port.
  • SolidWorks block:
    • Fixed a bug due to which the block sometimes did not close the SolidWorks application.
    • Fixed a bug due to which Simulation subtree could appear in the document tree several times.
    • Corrected document tree tooltip in case when simulation results were not available.
    • Fixed block’s behavior when it hid studies which were not static from the Simulation subtree of the document tree.
  • Text block: fixed incorrect error handling behavior when the block was unable to process an empty file.
  • GUI:
    • Fixed an error that occurred on closing pSeven while the GUI is still loading.
    • Fixed a GUI freeze when building a correlation plot.
  • General:
    • Fixed incorrect port name filtering in Workflow configuration dialog due to which the respective parameters could be removed from the Parameters and Monitoring panes.
    • Fixed an error due to which a GTDR model could not be exported in cases with different compressed vector dimension values.
    • Fixed issues which could sometimes result in pSeven deadlock when loading wrong models in pSeven Core blocks.
    • Fixed pSeven Agent crash under Windows with Anaconda installed.
    • Fixed various minor bugs in pSeven Agent.
    • Fixed the bug which resulted in the inability to re-open the report using a link in the Project description editor window.
    • Fixed the bug which resulted in the inability to use the Ctrl / hotkey for script editing in PythonScript, Text and Program blocks.
    • Fixed various problems with upgrading files from future versions.
    • Fixed the bug when the running workflow sometimes could not be terminated by the user.
    • Fixed an issue with Cyrillic characters not being properly displayed in the name of the dedicated folder after pSeven installation is over.
    • Fixed the encoding issue when installing pSeven under Windows.
    • Fixed the bug due to which pSeven installer was unable to find a license file in its default location.
    • Other minor bugfixes in pSeven installer.
    • Fixed problems with encoding of error messages that may sometimes appear in the console log when running the workflow.

pSeven 6.10

Release date: 11 January 2017.

This release includes pSeven Core (MACROS) 6.10.

New Features

  • New FloEFD block enables direct integration with FloEFD for SolidWorks (FloEFD standalone). See the FloEFD block page for details.
  • Analyze:
    • New approximation model export formats: executable file, Excel compatible DLL, and an Excel spreadsheet with a model DLL already integrated. Note that Excel export is supported only in pSeven for Windows.
    • Parallel coordinates and Sample viewer: added the Create new data series… command to viewer menus. This command creates copies of data series used by the current viewer and respects value filters applied in viewer configuration: new data series added to the report database contain only those values from the original data series that pass the filters.
    • Added the Text viewer that allows adding comments to reports. Supports CommonMark formatting.
    • Improved image support in the Page viewer. pSeven now automatically recognizes PNG and JPEG files generated by the workflow and stores them in the project database so they can be loaded into a Page viewer.
  • ApproxPlayer block: added the support for more export formats: C header, C source, C source for Microsoft Excel VBA, and FMU for Co-Simulation source.
  • Program block: remote execution options can now be sent to the remote port. Options received at the workflow run-time override those specified in block’s configuration.
  • General:
    • Enabled \(\TeX\) syntax support in project and workflow descriptions. Formulas should be enclosed in $$, for example $$y = x^2$$.
    • Added pop-up notifications for information messages, warnings and errors.
    • Added the project file association for Windows: double-clicking a .p7prj file now launches pSeven and opens this project.
    • Added a command line option to open a project on start (p7gui --project <path>).
    • Added a command line option to start pSeven without connecting to DATADVANCE for update checks and error reporting (p7gui --no-external-connections).

Updates and Changes

  • Edit:
    • Copy, cut, and paste functions now work with multiple blocks and annotations, and preserve links between blocks.
    • Added an option to select or deselect all types in port configuration dialogs.
    • Added a warning when the project directory is selected as some block’s sandbox (see section Sandbox).
  • Run:
    • Redesigned the parameter and monitoring panes and added the Statistics tab that is now used to show workflow run statistics, replacing the Workflow finished dialog.
  • Analyze:
    • Added commands for copying current dataset configuration to or from another existing dataset in viewer configuration dialogs.
    • Improved the display of error values in the Sample viewer.
    • Added an option to enable or disable table cell coloring in the Sample viewer.
    • Added the minimize command to viewer tiles.
    • Added keyboard hotkeys for various viewer commands.
    • A viewer tile can now be selected to put it in focus for keyboard hotkeys. Clicking an empty area in the report deselects the tile.
    • Enabled keyboard navigation in reports.
    • Added automatic scrolling when moving a tile to the top or bottom of the report.
    • Clicking an empty area in the Data series or Models pane now clears existing selection.
    • Added more information to model details.
    • Reduced memory usage when exporting models.
    • Improved color palette for plots.
  • Excel block: improved the handling of errors from Excel.
  • ANSYSWorkbench block:
    • Added an option to save changes to the project — see Save project.
    • Added an option to force unlock an incorrectly closed project — see Force project unlock for details.
  • Creo block: added an option to set the maximum number of block startups that can use the same Creo application instance.
  • Optimizer block: improved the support for NaN objective and constraint values and missing values in the initial designs data (the designs port). Such points are now taken into consideration if values of variables are within bounds defined by the problem. Also, Optimizer requests additional evaluations to fill the design points with missing values, and these evaluations do not consume the solving budget.
  • ApproxPlayer, DRCompressor, DRDecompressor blocks: reduced memory usage when exporting models.
  • Program block: if the program to run is located in the project directory, the block now stores a relative path to the program executable, unless you manually input an absolute path in its configuration.
  • Text block: added an option to select character encoding for the template, input and output files — see Text encoding.
  • General:
    • pSeven installer now supports installation for the current user only. Installing for all users requires administrator rights.
    • Various minor improvements in GUI.


  • New examples added to the pSeven example package — see them in Workspace:
    • 2.14 General - Excel Integration
    • 2.15 General - FloEFD Integration
    • 2.16 General - Multi-Objective Robust Optimization
    • 2.17 General - Parallelization Using Remote Hosts
  • Examples added or updated in the latest release are now marked with the #new tag. Also added tags that show pSeven version in which an example was added or updated.
  • Added detailed Text block documentation.
  • Added the FloEFD block page.
  • Updated section Options on the ANSYSWorkbench block page.
  • Updated section Document Tree on the SolidWorks block page.
  • Updated section Open Source Components.


  • Workspace: fixed an error when trying to view a non-Unicode text file.
  • Edit:
    • Fixed a crash when editing a PythonScript block with a lot of ports.
    • Fixed the inability to assign a value (default file) to a File port.
    • Fixed incorrect behavior after selecting an invalid or corrupted model file in ApproxPlayer, DFPlayer, DRCompressor, or DRDecompressor.
  • Run:
    • Fixed a crash when running a workflow that sends a big volume of data to its output.
    • Fixed incorrect workflow finish status being reported in case when the workflow is stopped or terminated.
  • Analyze:
    • Fixed pSeven freezing if you try to close it while a model is being trained in the Model builder.
    • Fixed incorrect Model validator behavior after some validated model is renamed on the Models pane.
    • Fixed the Models pane sometimes showing an incorrect technique during model training.
    • Fixed incorrect behavior when importing models created in some old versions of pSeven.
    • Fixed incorrect behavior of the Smooth model dialog after an attempt to input an incorrect smoothing factor.
    • Fixed incorrect behavior on the Correlations tab in the Sample viewer when deselecting dimensions in a multidimensional sample.
    • Fixed the Correlations tab in the Sample viewer not respecting the value filters specified in viewer configuration.
    • Fixed incorrect coordinates displayed in point tooltips in the Model explorer.
    • Fixed incorrect rendering of special symbols in axis labels in the Model explorer.
    • Fixed incorrect behavior of some 2D plots when zooming in and out.
    • Fixed a bug in automatic color selection on plots that sometimes resulted in showing different curves with the same color.
    • Fixed large tooltips sometimes being displayed out of the screen area.
  • Creo block:
    • Fixed a bug due to which the block failed to update model parameters with values received to input ports.
    • Fixed an error when building a document tree for a model containing section views.
    • Fixed a bug due to which the block could not regenerate the model that requires two or more regenerations.
  • Kompas3D block: fixed reopening the document on each startup when the document is a ZIP archive.
  • Composite block: fixed incorrect cache behavior (false hits) when cached inputs contain Dict values.
  • Optimizer block:
    • Fixed non-deterministic behavior of surrogate based optimization algorithms in multithreaded environment.
    • Fixed incorrect behavior when a floating point number is specified as an initial guess for an integer variable.
  • Program block: fixed the program or script output sometimes not being shown in the run log.
  • Formula block: fixed a memory leak.
  • General:
    • Fixed incorrect file path conversion in workflows transfered from Windows to Linux.
    • Fixed incorrect pSeven installation when the installer is run right after uninstalling the previous version without a reboot.
    • Fixed the possibility to crash pSeven by changing the system time while pSeven is running.
    • Fixed various minor bugs and inconsistencies in GUI.

pSeven 6.9 Service Pack 1

Release date: 3 November 2016.

This release includes pSeven Core (MACROS) 6.9 Service Pack 1.



  • Creo block: fixed a bug due to which the block sometimes did not close the Creo application correctly.
  • General: fixed incorrect behavior after using the File ‣ Save workflow as… command.

pSeven 6.9

Release date: 7 October 2016.

This release includes pSeven Core (MACROS) 6.9.


Licenses for pSeven 6.9 and above require the updated DATADVANCE vendor daemon which you can get from the license server package available at the DATADVANCE download page. The updated daemon is compatible with earlier versions of pSeven and the license server. For more details on the server and daemon setup, see Server Configuration.

New Features

  • New Excel block enables direct integration of Excel calculations into pSeven workflow thanks to the support for Excel documents with formulas and macros. See the Excel block page for details.
  • Predictive modeling tools:
    • Added the support for exporting a model as a Functional Mockup Unit for Co-Simulation (FMI standard).
    • Added Model explorer, a new tool for comparing models and studying model behavior. See section Model Explorer for details.
    • Added the support for calculating all model outputs and writing them to a cell range for the models exported to the Excel DLL format.
  • Edit: added the Duplicate command that creates a copy of the selected block.
  • Run:
    • When a workflow is running, individual blocks can now be interrupted manually, allowing certain real-time control over the workflow execution. See section Manual Stop for more details.
  • Analyze:
    • New tiling layout for reports that replaces old viewer windows with tiles that are snapped to grid, making viewer management much more convenient.
    • Added the Split data… command to create training and test data sets. See section Split Data for details.
  • Optimizer block: in addition to selecting an optimization preset, you can now specify which optimization methods to use when solving the problem. This is an advanced option that allows explicitly configuring the problem as single- or multi-objective, enables robust or surrogate based optimization, selects global search methods, and other. See GTOpt/Techniques for details.
  • General:
    • Added the Save project command to the main menu. This is a convenience command that simply saves all open workflows and reports.

Updates and Changes

  • Predictive modeling tools:
    • When evaluating models, you can now generate input data samples on the fly, in addition to the previously available option of using data from the report database as input. See section Making Predictions for details.
    • If the model includes submodels, the technique name in model details will now be "Composite", not "Auto".
    • Improved performance of models exported to the Excel DLL format.
    • Improved handling of categorical variables in SmartSelection algorithms.
    • Internal validation is now enabled by default when building approximation models, since in most cases it is needed for further model analysis in Model validator. Can be disabled using the “Internal validation” hint.
    • Enabled selecting error type (absolute or relative) when showing a scatter plot in Model validator.
  • Workspace: project description editor window can now be maximized.
  • Edit:
    • Regrouped blocks in the Block Library and renamed the ShellScript block to Program (see the updated Blocks section for details). Also, block names no longer include the “std” prefix. These changes are purely visual and do not affect compatibility with previous versions.
  • Analyze:
    • Removed partial correlation methods from the Sample viewer.
    • Removed the automatic report update function that caused various issues. Note also that the manual update button was moved to the main report toolbar since the view toolbar was removed due to the update in the report layout engine (the new tiling layout). Also, this button now animates when reloading is in progress.
  • Composite block: several Composite blocks can now use the same cache file, sharing the cached data. See section Cache for details.
  • Optimizer block:
    • Configuration now requires to select an optimization preset.
    • Improved results filtering in multi-objective robust optimization.
    • Added proper support for NaN objective and constraint values and missing values in the initial designs data.
    • Reworked the computational budget allocation policies in various optimization algorithms, making them more consistent and as a result increasing stability of affected optimization methods. This change primarily affects methods involving global optimization stages, such as the surrogate based optimization, that deal with computationally expensive problems (also including usage of these methods in robust optimization problems).
  • ApproxPlayer block: clarified warning messages related to smoothing.
  • Program block:
    • Improved command preview in the configuration dialog.
    • Added Unicode support in shell scripts and an option to save script files in Unicode.
  • GUI:
    • Corrected descriptions and control labels in predictive modeling tools and added more tooltips.
    • pSeven now remembers the order of workflow and report tabs after you close the project.
    • Various other usability improvements.
  • General:
    • The option to start pSeven after installation was removed from the pSeven installer since it can lead to unexpected behavior.


  • New examples added to the pSeven example package — see them in Workspace:
    • 2.08 General - Dimension Reduction - Feature Extraction
    • 2.11 General - Sample-Based Data Fusion
    • 2.12 General - Blackbox-Based Data Fusion
    • 2.13 General - Predictive Modeling
    • 3.6 Design Space Exploration - Three-Section Beam
  • Updated all existing examples, using new tiling layout in reports and fixing a few issues.
  • Restructured the Block Library and blocks documentation: all related information is now found in the Blocks section.
  • Added the Excel block page.
  • Updated GTOpt documentation in section pSeven Core:
  • Updated the Composite block page.
  • Added Model Explorer description and updated other sections in Predictive Modeling Tools.
  • Added the data split function description in section Split Data.

Compatibility Issues

  • This release includes pSeven Core 6.9 which applied significant changes to the computational budget allocation and selection policies in various optimization algorithms that are used in the Optimizer block. Due to these changes, in most optimization problems that include computationally expensive functions the solving path has changed as compared to the previous versions of pSeven and pSeven Core. This has lead to a difference in optimization results which can be observed when comparing results from pSeven 6.9 and earlier versions. This difference is not an error and should not raise any issues, since in expensive problem context the optimal solution can always be established only approximately so direct comparison is never reliable.

    This change does not introduce any compatibility issues with the workflows created in previous pSeven versions. However, you may want to review the Optimizer block configuration in tasks involving global optimization — that is, surrogate based optimization (problems that include expensive functions, see Hint Reference) or problems where a non-default GTOpt/GlobalPhaseIntensity value was used. In both these cases, a certain adjustment of GTOpt/GlobalPhaseIntensity may be needed; for more details on selecting the value to use, see section Local and Global Methods.

  • Changed the syntax of references to existing environment variables in environment settings in the Program and PythonScript blocks. Previously it used names of variables beginning with @ (@VARIABLE) — for example, to append a directory to system PATH:


    Since 6.9, the syntax is ${VARIABLE}, for example:


    This change causes an issue if you run a workflow from previous versions that contains a Program or PythonScript block with environment settings using the old syntax. To resolve the issue, you will have to manually update configuration of such blocks. Note that the ${VARIABLE} syntax is used both in Windows and Linux versions of pSeven; the Windows system syntax (%VARIABLE%) should not be used.


  • Workspace: fixed incorrect name conflict resolution when creating a copy of an existing directory on the Filesystem pane.
  • Edit:
    • Fixed the inability to use NaN values in port defaults.
    • Fixed port name sorting in the Links dialog.
    • Fixed incorrect handling of external hyperlinks in workflow annotations.
    • Fixed a bug due to which workflow annotations were ignored when grouping, and annotations inside a Composite block were removed on ungrouping it.
    • Fixed minor grid displacement due to which workflow elements were not snapped to the grid correctly.
    • Fixed synchronization errors that could occur when loading a model in ApproxPlayer, DFPlayer, DRCompressor, or DRDecompressor takes significant time.
  • Run:
    • Fixed block startup counters being reset to 0 after the workflow finishes. Also fixed counters for blocks nested into a Composite block.
    • Fixed empty run log for terminated workflows.
  • Analyze:
    • Fixed a bug that could lead to inability to save a report containing a model trained using SmartSelection.
    • Fixed SmartSelection not being able to train the model if given a test sample that contains only one point.
    • Fixed incorrect SmartSelection behavior in some tasks where the training sample includes categorical variables.
    • Fixed smoothing methods accepting NaN smoothness factor values.
    • Fixed incorrect export of model to C in case when the model has many independent outputs.
    • Fixed model details sometimes including option values that were not set by user (showing defaults).
    • Fixed slow loading of reports with heavy 3D plots which could crash pSeven if the CPU is slow enough.
    • Fixed a bug due to which model training was not aborted when the model is removed from the report database.
    • Fixed the “model list load error” in Model builder.
    • Corrected various validation messages in Model builder.
    • Fixed wrong indexing of inputs and outputs in Model builder.
    • Fixed a few interface issues in Model validator.
    • Fixed error in Sample viewer that appeared when a value filter was applied to a data set containing strings.
    • Sample viewer: fixed a bug in Kendall correlation due to which it could be less than -1.
    • Fixed an error when exporting sample statistics from the Sample viewer.
    • Fixed incorrect behavior of the Dependency tab in Sample viewer when the sample contains strings.
    • Fixed incorrect slider behavior in Page viewer.
    • Fixed the correlation threshold slider in Sample viewer displaying incorrect (rounded) values.
    • Fixed a bug in 3D plot viewer due to which it ignored the point display limit when using the triangulation method for surface reconstruction.
    • Fixed some validation messages in the Sample viewer.
    • Fixed the export command in the Page viewer that generated invalid HTML.
    • Fixed an error on removing a data set from the Page viewer.
    • Fixed incorrect rendering of 2D plots that contain a single point.
    • The Change source command in the report database is now correctly disabled for data series imported from files or when the project database is empty.
    • Fixed a bug which could cause a freeze when importing data from an Excel file.
    • Fixed pSeven sometimes marking reports as changed right after opening them.
  • Composite block: fixed cache not ignoring unconnected output ports.
  • Creo block:
    • Fixed an error on opening a part file with an index number appended to file extension.
    • Fixed a malformed error message when opening an archive.
  • Optimizer block:
    • Fixed incorrect values of integer vector variables being generated when vector dimension is a workflow parameter.
    • Fixed incorrect behavior of the configuration dialog after an invalid tolerance value is input on the Robust optimization tab.
  • Program block:
    • Fixed incorrect validation of file names in block configuration.
    • Fixed a bug due to which the block could incorrectly consider a cluster job finished when working with a Torque cluster, caused by long job identifiers.
    • Fixed incorrect behavior when the block is stopped manually.
    • Corrected some warnings and validation messages.
  • PythonScript block: fixed a bug that caused slowdown in long-running workflows containing a Composite block with a nested PythonScript.
  • Text block:
    • Fixed a Unicode decoding error in template files.
    • Fixed duplicated error messages in the configuration dialog.
  • GUI: various fixes of interface layout and behavior, typos in warning and validation messages.
  • General:
    • Fixed incorrect behavior of file ports when a block outputs default values in case of error.
    • Fixed a bug in project upgrade due to which it could fail on projects containing backups.
    • Fixed pSeven example projects containing backups from previous versions.
    • Fixed a bug that allowed to start pSeven Agent on the same host more than once, causing various errors.
    • Fixed incorrect behavior of code editor in scripting blocks in some cases.
    • Fixed incorrect behavior of data editors when current data type is changed.
    • Fixed pSeven not accepting lowercase drive names in file paths under Windows.
    • Fixed the incompatibility of pSeven licensing system with newest Linux distributions.
    • Fixed incompatibility with license files for older versions that appeared in pSeven 6.8.
    • Fixed the msvcr120.dll error when the local license server is started manually.
    • Fixed pSeven console log not working when the license file is placed into the installation directory.
    • Fixed pSeven not deleting its temporary files in some cases.
    • Fixed pSeven batch interface (p7batch --run) not working in CentOS.
    • Fixed missing or malformed console output when using the pSeven batch interface in Windows.
    • Fixed incorrect behavior of the pSeven uninstaller when it is started by an unprivileged user.

pSeven 6.8

Release date: 20 July 2016.

This release includes pSeven Core (MACROS) 6.8.

New Features

  • Workspace: added search function to the example browser. Supports full text search in example descriptions and filtering by tags (that are now added to descriptions).
  • Run:
    • Added counters to blocks showing the number of block startups in the current workflow run.
    • A workflow can now be configured to overwrite existing project database records (the run history) using the overwrite command from the monitoring pane menu. Default behavior is to append data to existing records, same as in previous pSeven versions.
  • Analyze: the Model builder tool (see Model Training) now supports the SmartSelection technology — an intelligent model training algorithm that automatically selects and configures an approximation technique in order to obtain the most accurate model, based both on training data properties and your knowledge about the data and the underlying dependency that can be added to the model builder configuration using various training hints offered by pSeven. SmartSelection is now the default training mode, and when needed, you can switch to the manual configuration in the Model builder dialog.
  • Analyze: other improvements in predictive modeling tools.
    • Added options for handling non-numeric values in training samples to the model builder (and the ApproxBuilder block). Approximation models can now predict undefined function behavior: the model trained with GTApprox/OutputNanMode set to “predict” will evaluate to NaN in areas near those points of the training sample that contain NaN output values. Other settings can be used to automatically remove the data points with non-numeric values from the training sample, or to stop with an error if such points are found.
    • You can now queue model training from the Model builder dialog.
    • The training log shown in model details is now updated in real-time when training the model.
    • Added model retrain command: you can now repeat the training process starting with an existing model. This allows applying quick changes in training settings without reconfiguring the model builder each time — for example, to try different training options with the same data.
    • Added the support for additionally smoothing a model after training. Smoothing is also available for existing models from previous versions.
    • Model details now include information on SmartSelection hints specified when training and descriptive statistics for the sample data.
    • Improved performance of the High-Dimensional Approximation (HDA) technique. This update also applies to ApproxBuilder.
    • The Models pane in model validator now features an option to synchronize the selection of plotted output between all loaded models. When enabled, changing the displayed output of any model automatically switches other models to the same output.
    • Model validator can plot both absolute and relative model errors.
    • You can now change location of the plot legend in model validator, or even disable the legend.
  • Analyze: updated Sample Viewer.
    • Added background coloring on the Data tab to show high and low values in each column.
    • Error values are now shown on the Data tab if errors were added to the dataset.
    • Added an option to filter shown points by index in dataset configuration.
  • Analyze: added a command to empty data series in a report database (discard their data values) without deleting the data series themselves. This feature is useful when creating preconfigured reports intended to be updated after running a workflow — like the reports found in many project examples installed with pSeven.
  • Program: added the support for executing applications remotely on a Windows host using the pSeven Agent — see section pSeven Agent for details.
  • Composite: added the support for handling errors from nested blocks — see section Error Handling for details.

Updates and Changes

  • Workspace:
    • Workflows and reports can now be opened without manually opening their project first.
    • To avoid visual clutter, the backup files created when pSeven upgrades a project from an older version are now stored in a separate subdirectory inside the project (this subdirectory is hidden by default).
  • Edit:
    • pSeven will now keep the order of workflow tabs when reopening a project; tabs are no longer sorted alphabetically by name.
    • The tab with environment settings was removed from configuration dialogs of most blocks; configuring environment variables is now supported only in the Program and PythonScript blocks. Also you can now use values of existing environment variables when defining new ones.
    • You can now use arrow keys to move blocks in the workflow (hold Shift for precise movement).
    • Added more keyboard hotkeys. Key combinations can be found in tooltips.
  • Analyze:
    • Predictive modeling tools:
      • The names of data series used for training are now saved in model information and are further used as names of model input and outputs — for example, when adding it to model validator.
      • An initial model for incremental training (supported by the GBRT technique) can now be selected in the model builder dialog.
      • After a model is evaluated in the Make predictions dialog, pSeven can now automatically open a sample viewer showing inputs and results.
      • Model information shown on the Parameters tab in model details dialog now includes a summary of features available in the model (like smoothing and accuracy evaluation support) and more detailed information on options used when training.
      • Added the Annotations tab to model details dialog. This tab shows comments stored inside the model, if any.
      • Model validator now supports a separate test sample for each loaded model. Due to this, the Test sample tab was removed from the configuration dialog.
      • The RRMS (relative root-mean-square) error metric in model validator is replaced with the more common \(R^2\) metric (coefficient of determination).
      • Many minor usability improvements and fixes.
    • Added color axis support to 2D plots.
    • Default color palette for plots now excludes colors that are too similar to each other, for better usability.
    • Disabled non-linear scales for the vertical axis on column charts.
    • Changed the default naming scheme of data series to avoid long names when possible.
    • Removed the data series rename command from viewer configuration dialogs. Data series can be renamed from the Report database pane.
  • ANSYSWorkbench:
    • Relative paths to ANSYS Workbench projects are now supported.
    • Parameter descriptions are now shown on the Project tree pane.
  • Optimizer:
    • Improved solving algorithm for mixed-integer surrogate based optimization (SBO) problems — the points to evaluate that are generated at later solving stages are now more evenly distributed in the design space.
    • Narrowed the selection of optimum solutions in robust optimization problems to avoid adding dominated points into the optimal set.
  • ApproxBuilder: added the Table Function technique (simple table lookup). See section Table Function for details.
  • Program:
    • Program output is now stored on disk instead of keeping it in block memory in order to optimize memory usage.
    • Added the support for inserting text and file substitutions into the “Program” (command line) field and mapping substitutions to workflow parameters.
    • The block can now monitor the standard output and error streams for specified text or regular expression and raise an error when they are found.
    • Environment settings specified in block configuration now also apply to the remote host in case of a remote execution.
  • PythonScript: the block now raises an error if an output variable was not initialized during execution, instead of outputting a None value. This error is processed as specified by error handling settings, which are also supported in PythonScript since this version. See section Error Handling for more details.
  • Text: the block now allows specifying a relative path to the template file and does not automatically convert it to an absolute path.
  • GUI:
    • Generally higher stability, better responsiveness, and faster startup thanks to the new rendering engine. Note that this update introduced a few more graphics requirements to the system where pSeven is installed, such as the True Color display for Windows or libxcb and libGL requirements for Linux — see System Requirements for details.
    • Various minor usability improvements.
  • General:
    • pSeven now stores all its files (examples, settings, projects) in a directory named pSeven, located in your documents directory. Note that a version number is no longer added to the directory’s name, unlike previous releases.
    • pSeven installation now includes two executables, p7gui (the default graphical interface) and p7batch (the batch command-line interface). The executables are found in the client subdirectory of the pSeven installation directory.
    • The interactive command-line interface formerly known as the pSeven Shell is no longer supported. Scripts can be run using the batch interface (p7batch --run).


  • Updated section System Requirements.
  • Added section Unattended Installation.
  • Added section Advanced Settings on the Block Configuration page. This section describes advanced block configuration features such as caching, error handling, and other.
  • Added section pSeven Runner that explains the usage of this special pSeven mode.
  • Updated the SolidWorks block documentation page.
  • Updated the Program block documentation page, added section pSeven Agent.
  • Updated the PythonScript block documentation page.
  • Updated and reworked pSeven examples. Each example is now a separate project in order to better show the capabilities and features of pSeven.
  • Added a MATLAB integration example — see 2.9 General - MATLAB Integration in the example browser in Workspace.


  • Workspace:
    • Fixed a bug in the project function that could lead to an error when a workflow from an older version is added to a project of the current version.
    • Fixed incorrect parsing of command line arguments when running a Python script.
  • Edit:
    • Fixed two ports being created when uplinking a port that has a comma in its name.
    • Fixed keyboard hotkeys not working after editing an annotation.
  • Run: fixed pSeven resetting the selection of sources in the run log when launching a workflow.
  • Analyze:
    • Predictive modeling tools:
      • Added the missing C header model export format.
      • Fixed the inability to show model details for some models.
      • Fixed the incorrect behavior of the value filter on the Training set tab in the Model builder dialog.
      • Fixed the ability to select an invalid model in the Make predictions dialog, leading to errors when pSeven tries to evaluate the model.
      • Fixed an error on loading models from older versions of pSeven in model validator.
      • Fixed the incorrect behavior of the Add model dialog in model validator.
      • Fixed the ability to save model validator configuration without selecting a model or data sample.
    • Fixed incorrect Sample Viewer behavior when viewing text data.
    • Fixed the point display limit working incorrectly on the Correlations tab in the Sample Viewer.
    • Fixed incorrect behavior of the plot zoom function in the Sample Viewer when the source data includes data series of different length.
    • Fixed incorrect Page Viewer behavior when displaying HTML data, Boolean values, and in some other cases.
    • Fixed incorrect behavior of the data export function in various cases.
    • Corrected window placement when the cascade layout is used.
    • Fixed reports being shown as changed right after reopening a project.
    • Fixed reports not being shown as changed when they are refreshed and new data is actually added to the report from the project database.
    • Fixed a warning in the pSeven console after duplicating a column chart.
  • NX: fixed the block appending the NX DLL path to the system PATH on each startup and finally crashing the workflow with the message “the environment variable is longer than 32767 bytes”.
  • Optimizer:
    • Fixed a synchronization error when setting a negative tolerance value in the robust problem configuration.
    • Fixed a bug in DoE generation at the model training stage in multi-objective surrogate based optimization (SBO) problems due to which the generated DoE did not fill the design space properly if problem variables had significantly different scales.
    • Fixed a bug in multi-objective surrogate based optimization due to which it did not count NaN results as performed evaluations, consequently exceeding the GTOpt/MaximumExpensiveIterations budget.
    • Fixed incorrect behavior (exceeding the budget) of the robust surrogate based optimization algorithm when NaN values are found in results of evaluations performed at the model training stage.
    • Fixed a bug in surrogate based optimization (SBO) due to which it could crash when training an internal approximation model.
    • Fixed the possibility of repeating evaluations in the global optimization mode.
    • Fixed a rare crash that could occur when problem solving is finished.
  • ApproxBuilder: fixed incorrect calculation of internal validation errors for models trained with the TA, iTA, or TGP technique in component-wise mode.
  • ApproxPlayer: fixed a hangup when loading a big model (such as a tensor approximation model) in the block’s configuration dialog.
  • DFPlayer:
    • Fixed the block not loading model details correctly.
    • Fixed issues related to switching between the sample-based and blackbox-based modes.
  • Program:
    • Fixed the block not supporting encrypted private SSH keys.
    • Fixed incorrect exit code when the block stops by timeout.
    • Fixed a warning about failing to clean the temporary working directory under Windows.
  • PythonScript:
    • Fixed the indent and unindent functions in the built-in code editor.
    • Fixed the invalid encoding of error messages under Windows.
  • Text: fixed an error when a directory is specified instead of a template file.
  • CSVParser: fixed an error when loading a CSV file exported from Analyze.
  • Const: fixed a memory overflow error when creating an integer constant with a too big value.
  • GUI: various minor fixes of interface layout and behavior.
  • General:
    • Removed the redundant do and done ports from workflow root blocks so they are no longer visible on the Inputs and Outputs panes in Run.
    • Fixed incorrect handling of relative paths and paths to files outside of the project directory in several dialogs.
    • Fixed a hangup in the file save dialog when trying to navigate to a prohibited location.
    • Fixed various issues with CommonMark text formatting.
    • Fixed a potential memory leak.
    • Fixed pSeven splash screen hiding the message box about a license error on start.

pSeven 6.7

Release date: 29 April 2016.

This release includes pSeven Core (MACROS) 6.7.

New Features

  • New ANSYSWorkbench block enables integrating ANSYS Workbench projects into pSeven workflows. The block supports mapping project input and output parameters to ports, allowing you to apply optimization and other design space exploration techniques to ANSYS Workbench projects. An example showing ANSYSWorkbench usage is also added to the pSeven examples package — see MEMS Optimization in the example browser in Workspace and section Examples.
  • SolidWorks: added SolidWorks 2016 support.
  • Run: added a tree-like view of workflow parameters on the Configuration tab. The tree interprets a dot as a name separator and automatically groups names with common parts.
  • Analyze:
    • Added the support for exporting approximation models to a special format that provides better compatibility with Excel. It exports model C code ready to be compiled into a DLL that can be imported in Excel without making changes in the code manually.
    • The Model builder uses significantly less memory when training high-dimensional RSM and HDA models. This update also applies to ApproxPlayer.
    • Redesigned 2D plots with better rendering, more visualization settings and multiple other improvements.

Updates and Changes

  • NX: the block now always validates the model after it is rebuilt in NX. This resolves the problem when NX returns success despite there were errors when rebuilding the model.
  • Composite: composite blocks now have common signal ports do and done. The behavior is the same as for other blocks - for example, the do signal is required to start if this port is connected (see section Signal Ports for details).
  • Const: allowed Unicode symbols in names of constants.
  • Run:
    • Block sandboxes are now cleared just before the block starts, not immediately after the block finishes execution. If the workflow finishes normally, pSeven still clears those sandboxes that should be cleared according to the sandbox management rule set on the Sandbox tab in block configuration. No sandboxes are cleared if the workflow stops due to an error.
    • All blocks that provide the “Error handling behaviour” option will now write a warning to the console log when they encounter an error but suppress it in order to continue workflow execution (the behavior is set to something other than “stop workflow”).
  • Analyze:
    • Updated and clarified many error and warning messages that are shown on the Issues pane.
    • The Scatter matrix tab in Sample Viewer is renamed to Correlations.
    • Added more hints on the Dependency tab in Sample Viewer.
  • GUI: minor user interface updates and usability improvements in various dialogs and data editors.
  • General:
    • Saving a workflow or report is now enabled even if there were no changes since opening them.
    • Added more keyboard hotkeys. Key combinations can be found in tooltips.



  • Optimizer: fixed a bug in the surrogate based optimization algorithm which could cause to it to work indefinitely in some problems.
  • ApproxPlayer: fixed incorrect smoothing of RSM models trained using the ElasticNet feature selection algorithm.
  • PythonScript: fixed failing validation of PythonScript blocks in workflows upgraded from version 5.2 and below.
  • Text: fixed wrong quick selection behavior on some numbers using the FORTRAN format.
  • Workspace:
    • Fixed the Ctrl Home hotkey not working in the Filesystem pane.
    • Corrected error messages when pSeven cannot create a new file, directory, or project.
  • Run:
    • Fixed an error in the duplicate preset command when the same preset is duplicated twice.
    • Fixed incorrect behavior after a workflow validation failure.
    • Fixed the terminate workflow function not working sometimes when the running workflow is in abnormal state.
  • Analyze:
    • The Model builder will now correctly handle very high point weight values which previously could result in training invalid models when GBRT, iTA, HDA, HDAGP, or RSM technique is used. This fix also applies to ApproxBuilder.
    • Fixed Sample Viewer incorrectly calculating statistics for data that includes integer values.
    • Fixed a bug due to which opening the Sample Viewer configuration dialog resetted the correlation and p-value thresholds set on the Correlations tab.
    • Fixed a hangup that could happen when manually refreshing a report that contains a Page Viewer.
    • Fixed a memory leak and possible GUI slowdown when all nodes on the Project database are expanded and there is a workflow running.
  • GUI: various minor fixes of interface layout and behavior.
  • General:
    • Fixed a bug that could result in a GUI freeze in long-running workflows.
    • Fixed incorrect error message sometimes being shown when the license limit is exceeded.
    • Fixed incorrect behavior of the file dialog when creating a hidden file or directory while showing hidden files is disabled.
    • Fixed various misprints and typos in error and info messages.

pSeven 6.6

Release date: 1 April 2016.

This release includes MACROS 6.6.


This release discontinues the support for Windows XP and other XP-based editions, as well as the support for some old Linux distributions. See the updated System Requirements section for full details.

New Features

  • Analyze:
    • Added predictive modeling tools to train approximation models, validate these models, evaluate models to get predictions, and export trained models to various formats. See section Predictive Modeling Tools for a full guide.
    • The Data series pane is replaced with the Report database pane that shows report data and models trained with the new Model builder tool.
    • Improved Excel import and export with better preview, more convenient settings interface and other usability improvements. Also, Excel import and export are now available as separate commands from the Report database pane menu.
    • Report database is now automatically defragmented and compacted when you save the report in order to reduce the database file size.
  • Added new examples showing the capabilities of pSeven’s robust optimization method, Amesim integration and features available in run-ready workflows. See Bracket Structure Optimization, Single-Cylinder Engine Optimization, and Run-Ready Workflows in the example browser in Workspace.
  • Optimizer: the block now uses a new robust optimization algorithm for problems that include expensive objectives or constraints. The new method is aimed to significantly reduce the required number of function evaluations at the cost of increasing the computational time spent in solving internal subproblems. Note that this overhead can be significant, since the method is intended for cases where function evaluations are time consuming or limited in number.
  • ApproxBuilder: added the support for incremental model training, sample weighting and using output noise variance data. See ApproxBuilder documentation page for details.
  • Creo: the block now provides full support for integration with PTC Creo 3.0.
  • PythonScript: simplified importing third-party modules, see section Python Modules on the block’s page for details.
  • Significantly reduced memory consumption for long-running workflows under Windows.
  • Workspace: new file browser makes working with large projects faster and more convenient.

Updates and Changes

  • Optimizer:
    • Improved result filtering — now the result excludes solutions that are insufficiently close to optimum even if they satisfy problem constraints. Due to this the number of points in optimization result can decrease compared to previous versions, but the result will contain more high-quality points.
    • Points in the initial designs sample for which values of variables are outside of variable bounds are now accepted and used, in particular, to build internal approximation models thus increasing model accuracy and stability. Previously such points were automatically removed from the initial sample.
    • The “Linearity type” setting for objectives and constraints is no longer ignored in robust optimization problems. However, Optimizer now assumes that functions specified as linear or quadratic do not depend on any stochastic variable. Note that it can lead to unexpected behavior in some problems with invalid formulation that were solved in previous versions of pSeven.
  • SolidWorks: improved error handling and block stability.
  • Composite: if one of the instances of a parallel Composite block crashes, it will now stop the workflow immediately (previously pSeven waited until all running instances finish).
  • PythonScript: the Variables pane can now be collapsed by double-clicking the pane border.
  • Program: the right pane showing substitutions and files can now be collapsed by double-clicking the pane border.
  • Submatrix: better hint text in the Add submatrix dialog.
  • Workspace:
    • The Project pane now uses list view similar to the Filesystem pane, instead of the tree view.
    • The Open command is now disabled for workflows and reports that are located inside some subdirectory of the project directory. Note that pSeven requires all workflows and reports to be located at the top level of the project directory since version 5.2.
    • Added correct error message when trying to copy an existing directory into self.
  • Edit:
    • Added visual notifications on long editing operations.
    • Disabled the Ungroup command for Composite blocks with enabled parallel execution (ungrouping such a block always breaks the workflow).
    • When you begin editing an alias or description in the Workflow configuration dialog, the input field will now be filled with default text initially. It simplifies setting custom aliases and descriptions which often are just slightly modified default values.
  • Run: clicking the edit icon on a workflow output now opens the Edit value dialog instead of starting inline editing.
  • Analyze:
    • Added commands to change the order of exported columns in data export dialogs.
    • The order of adding data series to a report database is now the same as the order of selecting project database records on the Project database pane.
    • Improved rendering of 3D plots.
    • Viewers will no longer ask for confirmation when you rename source data series from the viewer configuration dialog.
    • pSeven will now suppress unneeded warnings when importing CSV files with NaN or infinity values.
    • Disabled the tile layout for viewer windows.
  • General:
    • pSeven will no longer automatically escape backslashes in file paths and will not require this.
    • Data editors can now display HTML-formatted text.
    • Various small GUI improvements.


Compatibility Issues

  • Potential issue is the change in treatment of the “Linearity type” setting for objectives and constraints in robust optimization problems. Prior to 6.6, the algorithm to solve robust optimization problems implemented in the Optimizer block effectively ignored the linearity type setting for objectives and constraints in robust optimization problems, thus losing the opportunity to speed up evaluations of such functions. Version 6.6 improves robust optimization algorithms, in particular allowing linear or quadratic functions in robust optimization problems. The algorithm now does not ignore function type but assumes that functions specified as linear or quadratic do not depend on any stochastic variable (which is correct by definition). However, Optimizer currently does not check your problem formulation in this case. This can lead to unexpected behavior if you define a function that does depend on a stochastic variable, and at the same time you specify it to be linear or quadratic using the hint.

    Note that such problem formulation is incorrect and should be avoided for general reasons. In previous versions it was not an issue because function type was simply ignored in robust optimization problems.

    It is recommended to recheck configuration of Optimizer blocks in workflows from previous versions of pSeven if:

    • the workflow solved an optimization problem with stochastic variables, and
    • linearity type was set to linear or quadratic for some objective or constraint.

    In this case, you should verify that linear and quadratic problem functions do not actually depend on a random (stochastic) variable.

  • This release discontinues the support for the old version of the PythonScript block that was available only in pSeven 2.0 and below. Workflows containing the old block version can be opened, but will require certain updates done manually in order to return them to working state. For details, see section Old Block Versions on the PythonScript block page.


  • NX:
    • Fixed the ability to set NaN values for dimensions and expressions, which resulted into a corrupted exported model.
    • Fixed incorrect error handling when a part of an assembly is not found in the archive.
    • Fixed the block not working under certain system environment settings.
    • Fixed typos in license error messages.
    • Error messages from NX 10.0 and above will no longer be truncated. Messages from NX 9.0 and below are still truncated and could not be fixed, see section NX_issues on the NX documentation page.
  • Optimizer: fixed a bug which could cause incorrect initial sample generation in multi-objective and surrogate based optimization algorithms.
  • ApproxPlayer:
    • Fixed incorrect validation of the smoothness factor value.
    • Fixed the incompatibility of models exported to M-file format with older versions of GNU Octave (3.0.5 and below).
  • Text: fixed several minor usability issues.
  • PythonScript: fixed the inability to import some of the default Python modules provided by pSeven.
  • Program:
    • Fixed an issue with non-unique file names that caused conflicts when the block runs inside a parallel Composite block and connects to remote host or a cluster.
    • Fixed the block sometimes generating temporary output files with the same name, causing workflow errors.
    • Fixed the block sometimes not removing its temporary files under Windows.
  • UQ:
    • Fixed incorrect behavior of the block configuration dialog after removing all variables.
    • Fixed incorrect behavior of the Deterministic option.
    • Fixed the ability to set None and NaN values in correlation matrix and blackbox outputs.
  • Random: fixed incorrect behavior of the Deterministic option.
  • TextGenerator, TextParser (deprecated blocks): fixed the inability to reopen block configuration dialog after trying to preview template operations if there are no operations added yet.
  • Workspace:
    • Fixed errors and incorrect behavior of the file browser when opening folders that contain many nested folders and files.
    • Fixed incorrect behavior of the example browser window in some cases.
    • Fixed Text viewer ignoring the line ending setting.
    • Fixed incorrect modification time being displayed in file tooltips.
    • Fixed a slowdown when opening large projects.
  • Edit:
    • Fixed a slowdown when selecting a block with a lot of ports.
    • Fixed a synchronization error when selecting a block from the Workflow tree pane by Enter.
    • Removed incorrect validation warning about an ignored parameter value for output ports.
    • Fixed incorrect behavior of the workflow editor after ungrouping an Composite block.
    • Fixed typing a whitespace when editing a workflow annotation always inserting two spaces.
    • Fixed incorrect rendering of scrollbars in small annotations.
  • Run:
    • Fixed workflow output value not being shown if a data editor for this value was opened during workflow run.
    • Fixed incorrect display of workflow outputs that receive no value.
    • Fixed incorrect grouping on the Inputs and Outputs panes.
  • Analyze:
    • Fixed incorrect pSeven behavior when deleting many records from the project database at once.
    • Fixed errors on updating plots that have a large (about 100) number of data sources.
    • Fixed incorrect behavior of some 2D plots after zooming.
    • Fixed several usability issues in parallel coordinates plots.
    • Fixed empty viewers being displayed after maximizing, hiding, and then showing all viewers.
    • Fixed incorrect behavior of plots with large legends.
    • Fixed export commands being available in an empty Page viewer.
    • Fixed Sample viewer scrolling to the top of a table after changing column width.
    • Fixed the summary statistics on the Statistics tab in the Sample viewer.
    • Fixed incorrect behavior of the rename command in newly created (empty) datasets in viewer configuration dialogs.
    • Fixed errors when exporting Unicode string data to Excel.
    • Fixed data import and export dialogs sometimes losing keyboard focus.
    • Fixed incorrect import of CSV files containing NaN values under Windows.
    • Fixed a warning on importing data from a CSV file if multiple delimiter characters are specified in the import configuration.
    • Correctly prohibited renaming the special “[index]” data series in viewer configuration dialogs.
    • Fixed automatic refresh activating even when there is no workflow running and thus there is nothing to update.
    • Fixed indexing of transposed data series in the report database starting from 1 instead of 0.
    • Fixed indexing of data series imported from a file starting from 1 instead of 0.
    • Fixed a possible crash when closing a project with many unsaved reports.
  • GUI:
    • Fixed incorrect ordering of option lists on the Configuration tab in some block configuration dialogs.
    • Fixed the help button not working in some block configuration dialogs.
    • Fixed inconvenient scrolling in small tables.
    • Fixed pSeven sometimes showing an empty table cell after double-clicking it.
    • Fixed various typos in tooltips and messages.
    • Minor fixes of GUI layout and various text labels.
  • General:
    • Fixed errors on pSeven startup caused by outdated or broken program configuration files.
    • Fixed some project editing commands being available in pSeven Runner.
    • Fixed incorrect behavior of project and file open dialogs in several cases.
    • Fixed an error on closing a project or a report while report data is still loading.
    • Fixed string data editors and viewers sometimes not detecting HTML-formatted text.
    • Fixed the Static Mixer Optimization example project opening all workflows by default.
    • Fixed some minor memory leaks that occurred when closing block configuration dialogs or closing a project with several open reports.
    • Corrected warning messages when opening a project, workflow or report that have unsupported version.

pSeven 6.5 Service Pack 1

Release date: 30 December 2015.

This release includes MACROS 6.5 Service Pack 1.

Updates and Changes

  • Text: improved heuristics for determining field delimiters in templates.


  • Reverted the update of section GTDoE that was not yet ready to publish.


  • Text: fixed quick selection working incorrectly with template lines that contain trailing spaces — for example, ANSYS APDL files.
  • General: fixed a hangup on creating a new project that could occur if the last edited project contained an Text block.

pSeven 6.5

Release date: 23 December 2015.

This release includes MACROS 6.5.

New Features

  • SolidWorks:
    • Added SolidWorks 2015 support.
    • Added the support for static studies in SolidWorks Simulation. Supported input parameters are force, pressure, and torque. Supported output parameters are minimum and maximum stress and strain for elements and nodes, and minimum and maximum displacement for nodes.
  • Run:
    • Added a tree-like view on the Inputs and Outputs panes. The tree interprets a dot as a name separator and automatically groups names with common parts.
    • Reduced memory consumption and CPU load when running workflows that transfer big vectors or matrices between blocks.
  • Analyze:
    • Added Excel support. Data import and export in Analyze now works with Excel files.
    • Data import and export dialogs now support rectangular selections (a specific row range from selected columns) and allow selecting table rows.
    • Time-consuming operations in the sample viewer, such as training an approximation model on the Dependency tab, no longer block the viewer’s configuration.
    • When adding data from the project database to a report, pSeven now shows a warning if the data structure is ambiguous and provides a dialog where you can cancel the operation or transpose the data before adding.

Updates and Changes

  • Optimizer:
    • The internal limit on the maximum number of expensive function evaluations will no longer override GTOpt/MaximumExpensiveIterations if the latter is set to a higher value.
    • Added several internal changes in the multi-objective and surrogate based optimization algorithms that improve their stability in certain cases.
    • Changed the default solver preset to “Smooth problem”.
    • The “Noisy problem” and “Heavily noisy problem” presets now include predefined settings for GTOpt/NumDiffStepSize.
    • Updated and clarified the tooltips for solver presets.
  • ApproxBuilder: added the support for categorical variables to the piecewise-linear approximation (PLA) technique.
  • ApproxPlayer: added the support for exporting models trained using the PLA technique.
  • PythonScript: added the support for bokeh and matplotlib modules which can be used to create custom visualizations supported by the page viewer in Analyze.
  • Program: prohibited creating file and string substitutions with the same name.
  • Submatrix: added Unicode support in submatrix names.
  • Workspace: added a refresh button to the project description viewer.
  • Edit: the Uplink dialog now immediately validates the port names you specify in the “Uplinked to” column.
  • Analyze:
    • Removed redundant columns from the Axes panes in plot configuration dialogs.
    • The Change source dialog for data series will no longer allow selecting a source that cannot be formatted to data series.
    • Changed the name of the data series removal command from Delete to Remove.
    • Disabled non-linear value axis scales for stacked bar charts.
  • Analyze, sample viewer:
    • The “Manual bins count” setting is now disabled unless the histogram bins count calculation mode is set to “manual”.
    • Updated hints shown when no data is added to the viewer or there was a data processing error.
  • GUI: names of commands, panes, dialog windows are now written in sentence case.


  • Added section Categorical Variables that provides more details on training approximation models containing categorical (discrete) variables.
  • Added a sample viewer usage example to the General Examples project.
  • Fixed incorrect sample viewer settings in the example projects.
  • Updated section Sample Weighting.
  • Updated the GTApprox/CategoricalVariables option description.
  • Updated the GTOpt/MaximumExpensiveIterations option description.
  • Fixed incorrect embedded sRGB profiles in some images in the manual.


  • NX:
    • Fixed incorrect model export in case when the block works in a loop.
    • Fixed a logging error related to the NX DLL files path.
  • Optimizer: fixed a bug due to which a solution to a robust optimization problem or a problem with some expensive function (surrogate based optimization) could contain a non-optimal point if an initial sample containing values of variables and objectives is given.
  • Text:
    • Fixed quick selection working incorrectly with template lines that contain Unicode characters.
    • Corrected text of warning messages in the block configuration dialog.
  • Program: fixed the block requesting a Linux keyring password when testing an SSH connection.
  • PythonScript: fixed incorrect return type of the read() method of pSeven file-like object.
  • TextGenerator, TextParser (deprecated blocks): fixed an error on opening the block configuration window with the Operations pane collapsed.
  • Workspace:
    • Fixed pSeven not opening a project automatically when you try to open a report from this project.
    • Fixed the inability to open a HTML file if its name or path contains Unicode characters.
  • Edit:
    • Fixed errors that appeared in block configuration dialogs when reading large files — for example, big CAD models or part files.
    • Fixed incorrect behavior of the Paste command after copying a block or annotation to the clipboard and then closing the project.
    • Fixed incorrect behavior of the Import workflow dialog in case of wrong block name or workflow path.
    • Fixed pSeven not keeping parameter values of the Composite block that is exported to a workflow.
    • Fixed some GUI errors that appear when editing complex workflows.
    • Fixed a memory leak when opening and closing block configuration dialogs.
    • Fixed incorrect default workflow save path in some cases.
  • Analyze:
    • Fixed the Add to data series (transposed) command not being available while a workflow is running.
    • Fixed incorrect behavior of the autorefresh function after closing a report with autorefresh enabled.
    • Fixed a timeout when importing a big CSV file into a report.
    • Fixed column selection in the data import dialog being reset after you change some of the parser settings.
    • Fixed the inability to change the source of a data series from a port record to a single value record.
    • Fixed pSeven hangup when some data that cannot be formatted into data series is added from the project database to a report.
    • Fixed a background error on closing a project with an open report that is still loading.
    • Fixed a background error on closing a project with an unsaved report.
  • Analyze, sample viewer:
    • Fixed incorrect behavior of column selection on the Data tab when new data series are added to the viewer by drag-and-drop.
    • Fixed incorrect behavior of the Scatter matrix tab when the loaded dataset is 1-dimensional.
    • Fixed scrollbars missing from the Scatter matrix tab when the matrix does not fit into the viewer window.
    • Fixed incorrect handling of data samples with insufficient number of points on the Dependency tab.
    • Fixed a bug in updating the viewer configuration due to which new data series added by drag-and-drop were not available in data selectors on the Scatter matrix and Dependency tabs.
    • Fixed an error on renaming data series from the viewer’s configuration.
    • Fixed a background error when closing a sample viewer that still processes the source data.
    • Fixed non-escaped HTML in string values which could result in incorrect appearance of the Data tab.
    • Corrected some unclear validation warning messages.
  • Analyze, Page viewer:
    • Fixed an error on exporting a page that contains a single numeric value.
    • Fixed the unavailability of export commands after switching pages.
    • Fixed incorrect page count after switching datasets.
  • General:
    • Project upgrade now works correctly with locked projects and automatically removes the lock if the project is not open in another pSeven instance.
    • Fixed pSeven not closing sometimes when you click Yes in the exit dialog.
    • Fixed incorrect behavior when closing pSeven while there is a workflow running.
    • Fixed incorrect behavior of the pSeven installer when trying to install to a read-only location.
    • Fixed the Dict data editor sometimes generating non-unique key names.
  • GUI:
    • Fixed mouse scroll not working in text editors after removing focus from the editor.
    • Fixed mouse selection and scroll not working in the StringScalar data editor.
    • Fixed the vector, matrix and List data editors losing selection after you change the number of elements.
    • Fixed incorrect appearance of the Dict data editor after adding many key-value pairs.
    • Fixed a bug in hotkey handling due to which the exit dialog could appear twice when closing pSeven.
    • Corrected the warning message about a duplicate block name in the Rename block dialog.
    • Corrected name validation messages in various configuration dialogs.
    • Fixed non-escaped HTML in message boxes and the example browser.
    • Other minor fixes in GUI appearance.

pSeven 6.4

Release date: 27 November 2015.

This release includes MACROS 6.4.

New Features



  • Fixed the inability to rename data series on the Data series panel.
  • ApproxBuilder: fixed the GBRT technique crashing when the minimum weight of points in a leaf set by GTApprox/GBRTMinChildWeight is too high for the given (small) sample size.

pSeven 6.3

Release date: 19 November 2015.

This release includes MACROS 6.3.

New Features

  • ApproxBuilder:
    • Added new piecewise-linear approximation technique. See section Piecewise Linear Approximation for details.
    • Added an option to tolerate deviations of input values from a grid-like DoE, applying input rounding and allowing to use tensor approximation techniques with noisy samples with “almost factorial” DoE. See GTApprox/InputsTolerance and section Sample Cleanup (in particular, Input Rounding) for more details.
    • Added the support for discrete (categorical) variables in HDA, GP, SGP, HDAGP, TGP, and iTA techniques. Discrete variables are specified by GTApprox/CategoricalVariables; note that the same option now specifies discrete variables for the RSM and TA techniques, making the GTApprox/RSMCategoricalVariables and GTApprox/TADiscreteVariables options deprecated. The latter options are kept for version compatibility only and should not be used further, as they are going to be removed in future versions.
  • DoE:
    • Added the support for grid-based adaptive DoE generation. Setting GTDoE/CategoricalVariables now forces the adaptive generation algorithm to work on a user-defined grid, so the generated sample includes only grid points.
  • Program:
    • Added an option to keep temporary files (program input and output files) created by the block.
  • Analyze:
    • Added a new viewer type — the Page viewer, supporting plain text, CommonMark, and HTML.
    • When adding data from the project database to a report, existing data series are now re-used when possible, instead of creating new data series with an appended index number. In particular, re-adding data from the same project database record now works as a manual update of existing data series.
  • General:
    • Increased GUI responsiveness due to the support for asynchronous operations.
    • Added keyboard hotkeys for common functions and commands.
    • Text in error message boxes can now be selected and copied to the system clipboard.

Updates and Changes

  • Creo, Kompas3D, NX, SolidWorks: block configuration dialog now shows a brief help if there is no document loaded.
  • Optimizer: in the solution of a robust optimization problem, chance constraint values are now sorted after general (expectation) constraints.
  • ApproxBuilder, DFBuilder: componentwise approximation is now enabled by default in order to avoid problems when training models with many independent outputs. Old behavior (disabling componentwise approximation) can be achieved using the GTApprox/DependentOutputs option (or GTDF/DependentOutputs, respectively). Note that the GTApprox/Componentwise and GTDF/Componentwise options are now deprecated and are kept for version compatibility only. These options should not be used further, as they are going to be removed in future versions.
  • Program:
    • Added a 15 seconds timeout for SSH connections.
    • Allowed resizing panes in the block’s configuration dialog.
  • Text:
    • Operation configuration dialogs now show additional information when some controls are disabled due to operation’s custom settings.
    • Clarified and corrected error messages.
  • PythonScript:
    • Reduced memory consumption when working with big data arrays.
    • Added the support for requests, pandas, xlrd, and xlwt modules.
  • Workspace, Edit: project descriptions and workflow annotations will now use CommonMark — a more stable implementation of Markdown and its syntax.
  • Workspace:
    • Better responsiveness when viewing a project that contains many files and when viewing file tooltips.
    • Added a hint on the path to the default project location in the Create project dialog.
    • Added a close button to the title bar of the project description editor window.
    • Renamed the project creation button.
  • Edit:
    • Added navigation buttons to the [In] and [Out] bounds, to return to the upper workflow level when editing an Composite block.
    • Prohibited manually changing type of an Composite block port if this port is connected internally and exclusively (only to one link).
    • If you do not input a name for the new block in the Import workflow dialog, it now generates this name automatically based on the imported workflow filename.
  • Run:
    • Workflow outputs can now show text formatted with HTML or CommonMark.
    • When Enter is pressed in the Run workflow dialog, it now cancels running the workflow by default instead of starting the workflow. This is done to avoid accidentally running a workflow with validation warnings.
  • Analyze:
    • Compacted the report toolbar by gathering similar tools into groups.
    • Indices in default names of columns (data series) generated when importing data from file are now padded with zeros to allow correct name sorting.
    • Removed the option to smooth plot lines from the parallel coordinates plot configuration (smoothed plots in old projects will be preserved, but new smoothed plots can no longer be created).
  • Analyze, sample viewer:
    • You can now drag-and-drop data series into an existing sample viewer to add them as new columns.
    • Increased responsiveness when viewing large datasets.
    • Removed the multiple dataset support since it overcomplicates viewer configuration and is rarely needed.
    • Correlations on the Scatter matrix tab can now be calculated not only for the whole dataset, but for any subset of points. To recompute correlations, simply select an area on the plot.
    • The Scatter matrix tab now shows the number of selected points and the total number of plot points.
    • Renamed the Approximation tab to Dependency.
    • Errors plot on the Dependency tab now supports interactive zoom (select an area to zoom in, double-click the plot to reset zoom).
    • Clarified various GUI labels and messages, added more tooltips.
    • Removed manual approximation options input from viewer’s configuration.
  • Block configuration:
    • Block configuration dialogs no longer reset various pane and column sizes when you close the window.
    • Removed quotes and brackets around port type names on the Ports tab. Also if the port supports any data type, the “Type” column now shows “Any” instead of a full list of data types.
    • Unneeded quotes are no longer added around names of variables (when editing them) and in other input fields.
  • General:
    • User-defined names (blocks, ports, variables, monitor and parameter aliases, data series) are now limited to 256 characters.
    • Disabled editing project descriptions in pSeven Runner.
    • Changed the location of pSeven crash reports: they are now stored inside the pSeven settings directory (%USERPROFILE%\.p7\ in Windows, ~/.p7/ in Linux) instead of saving them to the desktop.
  • GUI:
    • Improved the formatting of floating point values in various tables and dialogs.
    • The current focused row in tables is no longer highlighted to avoid confusion with a selected row.
    • Input focus in the Add block dialog is now by default placed on the name input field instead of the OK button.
    • Various other minor improvements.


Compatibility Issues

  • Potential issue is the changed order of constraint values in Optimizer results. In prior versions, all constraints in results output by the Optimizer block were sorted in the order of adding them to the problem. This became inconvenient, if there are both general and chance constraints in robust optimization, since they are listed on different tabs and there is actually no way to recover in what order they have been added. Due to this, since 6.3 all chance constraints will be sorted after general constraints in optimization results, as described in section Robust Problem Solution on the Optimizer block page.

    This change does not break block-level compatibility (for example, ports and data types are not changed), though it can lead to certain data compatibility issues due to the changed order of columns. It is recommended to revise workflows that solve a robust optimization problem with chance constraints, if they somehow process the constraints data from the Optimizer solution.

    This change can possibly affect a workflow only if:

    • the workflow includes an Optimizer block, and
    • Optimizer solves a robust optimization problem, and
    • this problem has one or more chance constraints.


  • Optimizer:
    • Fixed fatal error when solving a mixed-integer problem without constraints.
    • Fixed incorrect behavior of the configuration dialog when using manual variable grouping.
  • ApproxBuilder: fixed a bug that sometimes lead to inability to train a tensor approximation model when model reduction (GTApprox/TAModelReductionRatio) is enabled.
  • PythonScript, Program, Text:
    • Fixed Ctrl Z clearing the script if it was not yet edited after opening the block configuration dialog.
    • Fixed incorrect behavior of tab indentation in the script editor.
  • PythonScript:
    • Fixed a version compatibility issue with workflows created in old pSeven versions.
    • Fixed a memory leak that was caused by repeatedly opening configuration of a block that defines many variables.
  • Program
    • Fixed a bug in the Windows version that sometimes caused the inability to remove the block’s sandbox if it contains executable files.
    • Fixed incorrect job status detection when working with the Torque job manager.
    • Fixed incorrect layouting of the configuration dialog window after resize.
  • Text:
    • Fixed the template view always scrolling to top when an error message appears.
    • Fixed removing the last new line from the end of the output file.
    • Fixed working with dictionary and list data types in the write operation.
    • Fixed the appearance of message boxes.
  • UQ: fixed a bug in block configuration due to which some workflows from the Robust Optimization example project could not start.
  • DoE: fixed incorrect behavior of the generation technique selector after switching the block working mode — for example, from batch to sequential.
  • IVE: fixed incorrect results in the sample-based mode if the analysis technique is not selected manually.
  • Submatrix: fixed incorrect row and column range validation in the Add submatrix dialog.
  • Formula (beta block): fixed incorrect behavior after adding a formula with an invalid name.
  • TextGenerator, TextParser (deprecated blocks): fixed incorrect behavior of the script editor.
  • Composite: fixed block description not being saved properly.
  • Condition: fixed a few incorrect validation warning messages.
  • Submatrix: removed empty columns on the Submatrices pane in the configuration dialog.
  • Block configuration:
    • Fixed incorrect input validation when assigning values to ports.
    • Fixed slowdown occurring in Integration blocks and some others when you open a block that defines many variables.
    • Fixed data editor dialogs interpreting numbers with leading zeros as octal and trying to convert them to decimal.
    • Fixed incorrect value validation for some options.
    • Fixed incorrect substitution of the default sandbox path on the Sandbox tab when the sandbox is set to project directory.
    • Corrected several error messages.
  • Workspace:
    • Fixed a crash when opening multiple workflows at once without opening the project first.
    • Project description editor: fixed Ctrl Z clearing all content if it was not yet changed after opening the editor.
    • Fixed the examples browser window blocking the dialog that informs about unsaved changes in the project, if trying to load an example while there is another project open.
    • Fixed example browser showing files and subdirectories in the Examples directory as projects.
    • Fixed project name validation in the Create project dialog.
    • Fixed pSeven not responding after opening and closing a text or CSV file, and after opening a HTML file from the Filesystem pane.
    • Fixed text formatting in various error message boxes.
  • Edit:
    • Fixed the inability to copy the Composite block you currently edit to the pSeven clipboard.
    • Fixed unclear error messages from the Import workflow dialog:
      • When trying to import a workflow with an unsupported version.
      • When the path to the workflow file is incorrect or not specified.
    • Fixed incorrect validation message when there is a port with assigned value and no link connected.
    • Removed the nonsensical warning about monitoring a port of the File type.
    • Fixed the Uplink dialog showing all blocks in the workflow if it is open while there is more than one block selected (should show only the selected blocks).
    • Fixed incorrect behavior of uplinks and corresponding ports when editing port types, if multiple uplinks are connected to the same port.
    • Fixed incorrect behavior of the Block library pane after expanding several nodes.
    • Fixed a noticeable slowdown of the workflow editor when making changes in a workflow with many blocks and links.
    • Fixed incorrect rendering of value tooltips in data editors in case when a value contains a HTML snippet.
    • Fixed orphaned block expand buttons that were sometimes shown in workflow editor after removing an Composite block.
    • Fixed copying text from annotations (copied the annotation itself instead of the text).
    • Fixed a GUI freeze when opening workflows or reports from another project while you already have a project open.
    • Fixed a GUI freeze if Cancel is clicked in a block configuration dialog before it finishes loading.
    • Fixed a synchronization error when many blocks are added to a workflow in rapid succession.
  • Run:
    • Fixed incorrect run name validation on the Configuration tab.
    • Fixed a bug which could cause a GUI freeze when the workflow outputs NaN values.
  • Analyze:
    • Fixed viewers sometimes not keeping their data sources after you rename a data source from the viewer configuration dialog.
    • Fixed the automatic selection of source data series of the currently active viewer on the Data series pane that did not work for the sample viewer in some cases.
    • Fixed data export not working when the data contains empty string values.
    • Fixed incorrect behavior of plot configuration when changing the order of axes.
    • Fixed the inability to update data series correctly if their source record has changed its shape.
    • Fixed a GUI freeze when opening workflows or reports from another project while you already have a project open.
    • Fixed pSeven sometimes not detecting changes in the edited report.
    • Fixed incorrect error message when the new name for a project record typed in the Rename dialog contains forbidden characters.
    • Fixed a crash when adding a record from the project database to report if the record contains a dictionary with an empty key or value.
    • Viewer configuration: fixed single click entering the cell edit mode on the Axes pane, General settings tab.
    • Fixed incorrect placement of big tooltips due to which they were unreadable.
    • Fixed rows numbered “-1” sometimes appearing in tables.
    • Fixed wrong aspect ratio on 3D plots.
    • Fixed incorrect sizing of 3D plot windows due to which most 3D plots were cropped at the bottom.
  • Analyze, sample viewer:
    • Fixed Scatter matrix tab sometimes not showing the warning about non-displayed points (exceeding the display limit).
    • Fixed scrollbars sometimes missing on the Scatter matrix tab.
    • Fixed an error occurring on the Dependency tab when the dataset contains many duplicates.
    • Fixed incorrect training of approximation model for a constant zero output and in some other specific cases.
    • Fixed incorrect behavior of the dependency input and output selectors.
  • Examples: fixed the Rotating Disk Optimization project not working under Linux if glibc 3.4.9 or newer is not installed.
  • General:
    • Fixed file open and save dialogs not remembering the last used path.
    • Fixed a crash when trying to save a workflow under a name containing 250 or more characters.
    • Fixed a crash when stopping the workflow right after start.
    • Fixed incorrect pSeven behavior when a project is closed before it is fully loaded (for example, when there are many opened workflows from the previous session).
    • Fixed Modeling blocks not showing error types correctly in some cases.
    • Fixed pSeven incorrectly determining workflow and report changed status in some cases.
    • Fixed the Issues pane toggle button remaining highlighted after the workflow or report that produced warnings is closed.
    • Fixed a memory leak occurring when there is a long log in pSeven console.
    • Fixed incorrect error message (“file exists”) from the file save dialog when trying to save to a location where you have no write access.
    • Fixed malformed error messages in some Integration blocks.
    • Fixed the font color for None values in data editors.
  • GUI:
    • Fixed some layouting issues in the Links dialog.
    • Fixed the inability to resize an error message box that contains additional details.
    • Fixed non-escaped HTML in message boxes that could corrupt the message text.
    • Fixed the unlimited resize of the Uplink dialog.
    • Various minor bugfixes.

pSeven 6.2

Release date: 13 October 2015.

This release includes MACROS 6.2.


This release does not provide 32-bit versions of pSeven due to lack of demand. pSeven continues to support 32-bit Windows and Linux systems, but only 64-bit installers are available for trial. Commercial 32-bit versions will be available on request (you can contact us for details).

New Features

  • NX:
    • If the loaded model is changed on disk, NX will now reload it automatically before building the document tree, allowing you to configure the block and edit the model simultaneously.
    • Improved the support for complex assemblies created using WAVE Geometry Linker.
  • ApproxBuilder:
    • Added an option to reduce the complexity of tensor approximation models — see GTApprox/TAModelReductionRatio.
    • Added the support for SGP (sparse Gaussian processes) factors to the Tensor Approximation technique.
  • Workspace: added the description edit button on the Project description tab (hover the text to show).

Updates and Changes

  • NX:

    • Better error handling, in particular when loading a model from a newer NX version on a system where an older NX version is installed.
    • The document tree will no longer show duplicate names of dimensions.
    • “Parameters” in the document tree are renamed to “attributes”.
  • Optimizer:

    • An improvement in the surrogate based optimization method enables solving problems with a lower limit for GTOpt/MaximumExpensiveIterations if an initial sample containing variable, objective, and constraint values is supplied.

      This update finalizes the improvement of SBO for large-scale optimization that was gradually implemented in previous pSeven versions and now includes:

      • The support for high dimensional problems with hundreds of design variables (see section Surrogate Based Optimization for details).
      • Noticeable runtime reduction thanks to hierarchical and multilevel surrogate based optimization.
      • Improved support for NaN responses (objective values).
    • Specific methods of handling cusp-like singularities in problem functions are no longer used if analytical gradients are enabled in the problem.

  • ApproxBuilder:

    • Significantly reduced size of models trained using the GBRT technique.
    • Reduced size of componentwise (GTApprox/Componentwise on) models trained using the GP technique.
    • Improved the algorithm that selects the subset of the training sample to be stored into a GBRT model. The stored subset is now more smoothly distributed over the initial training sample, which positively affects quality of incremental model training.
  • Workspace:

    • Opening a project now always switches to the Project description tab.
    • Added proper warnings if a project description is empty or corrupt.


  • Added NX block documentation.


  • Installer: fixed missing digital signing in the installer of the 64-bit version for Windows due to which it showed that the program comes from an unknown publisher.
  • NX:
    • Fixed export not working if the export file has port origin or the path to the export file is received to a port.
    • Fixed error message “License server system does not support this feature. [-18]” appearing in pSeven for Linux.
  • Optimizer: fixed a bug which could lead to incorrect behavior of the surrogate based optimization method in mixed-integer problems with small values of the GTOpt/MaximumExpensiveIterations option.
  • ApproxBuilder: fixed a bug in determining the GTApprox/HDAMultiMax bounds that did not allow values less than 5 (the default value for GTApprox/HDAMultiMin).
  • DoE: fixed a bug in the OLHS technique due to which it could generate a sample which is worse than a non-optimized LHS with the same random seed.
  • CSVParser: fixed a crash when previewing a CSV file under Windows.
  • CATIA (beta block):
    • Fixed handling the error that occurs when a locked parameter is mapped to a block variable.
    • Fixed incorrect error logging.
  • Workspace: fixed the inability to run a Python script in pSeven for Windows if the path to the script or its filename contain characters that are not found in the default codepage for the current system locale.
  • Analyze:
    • Fixed disappearing column selection on the Data tab in the sample viewer.
    • Fixed a possible GUI freeze when selecting multiple data series in the report database.
  • General: fixed several bugs in pSeven for Windows that appeared if the path to the project you are trying to work on contained characters that are not found in the default codepage for the current system locale.

pSeven 6.1

Release date: 21 August 2015.

This release includes MACROS 6.1.

New Features

  • Major improvements in NX:
    • Added the support for Siemens NX 10.
    • Added the support for Unicode file names and paths, names of model features and dimensions.
    • Expressions defined in the working model are now shown in the document tree.
    • The block now correctly handles the case when there are no solid bodies in the model and mass properties cannot be calculated.
  • Analyze: added a command to compact project databases to the Project database pane context menu. Note that a database can be compacted only if there is no workflow running and no reports are open.
  • ApproxBuilder: added the elastic net regularization method in the RSM approximation technique; see GTApprox/RSMFeatureSelection and GTApprox/RSMElasticNet/L1_ratio.

Updates and Changes

  • NX: the block now prohibits exporting a model with dependencies a part file format (.prt). Such models can be exported to a ZIP file as an assembly.
  • Optimizer:
    • Solving mixed-integer problems no longer requires using the surrogate based optimization (SBO) method. Previously to solve a mixed-integer problem it was required to define at least one objective or constraint as expensive.
    • Improved handling of cusp-like singularities in objective and constraint functions.
  • Creo: better handling of various cases when PTC Creo process does not finish for some reason.
  • Program:
    • When the block works over SSH and its sandbox is copied from a Windows to a Linux host, pSeven now sets executable bit for all files in the sandbox to avoid problems in scripts that run executable files from the sandbox.
    • Added more details to messages informing about the results of an SSH connection test.
  • PythonScript: enabled resizing the Variables pane.
  • Composite: when editing ports, their names and values are now validated immediately, so an invalid name or value does not bring up a message box.
  • Submatrix: enabled keyboard navigation in the Add submatrix dialog.
  • Block configuration: names of variables in the Environment tab are now validated immediately, and an invalid name does not bring up a message box.
  • Analyze:
    • The sample viewer will no longer fill shorter sample columns with None values in order to avoid confusion with missing values in the sample.
    • When importing data from a file with empty column names, pSeven will now insert default names in the unnamed columns.
    • 3D plots will no longer reduce the number of points shown if a 3D surface is not drawn on the plot.
    • The Data series pane now uses an inline editor to rename data series.
    • Disabled shadows under point markers on 2D plots.
  • General: added an option to toggle showing hidden files in pSeven file selection dialogs.
  • GUI:
    • Message boxes informing about errors now contain additional details.
    • Checkboxes in the Import workflow dialog now can be toggled by clicking the checkbox label.


  • The Russian version of the pSeven User Manual is now available from the Help menu.
  • Added SolidWorks block documentation.
  • Updated Tutorials.
  • General Examples project: added approximation model export to the sample-based approximation example, updated project description.
  • Updated and improved the Rotating Disk Optimization and Static Mixer Optimization example projects.
  • Updated the GTApprox/RSMFeatureSelection option description (added the elastic net regularization method).
  • Added the GTApprox/RSMElasticNet/L1_ratio option description.


  • SolidWorks:
    • Fixed incorrect parsing of global variables with negative values, due to which these variables appeared as read-only in block configuration.
    • Fixed occasional inability to delete temporary files created by the block.
  • NX:
    • Fixed a bug due to which the block could crash when used in a cycle with a large number of iterations.
    • Fixed dysfunctional model export to ZIP (changes were not applied to model before exporting).
    • Fixed the inability to export model to a part (.prt) file if the block works in a cycle. Export to other formats was not affected by this bug.
  • Creo:
    • Fixed a bug due to which the block did not close the working document file on finish if it connected to the Creo application already running in the background. This bug could lead to incorrect behavior when Creo receives the working file to a port.
    • Fixed “An error occurred while opening Creo document” message appearing when running a workflow that contains two or more blocks with different Creo application visibility settings.
    • Fixed changes to the Creo configuration file not being reverted after the block finishes.
    • Fixed occasional errors identified by messages “”solid” is referenced before assignment” and “”casync” is referenced before assignment”.
  • Integration blocks: fixed a bug in pSeven for Windows due to which the default value assigned to output ports was None instead on NaN. Due to this bug, if one of these blocks was used in an optimization cycle, its evaluation failures stopped the workflow. This is because by convention a NaN response value received by Optimizer indicates an evaluation failure (handled correctly), while None is not used and in general means absence of value (so-called missing values).
  • Program:
    • Fixed incorrect handling of placeholders in the names of files used to redirect standard input and output.
    • Fixed an error when printing a Unicode message to the pSeven console.
    • Fixed a bug due to which a block configured to connect to a cluster did not output stderr messages to the pSeven console.
  • Composite:
    • Fixed incorrect caching of missing values.
    • Fixed incorrect behavior in case when cache is enabled and there are unused output ports.
    • Fixed a bug in pSeven for Windows which caused a crash if the block’s cache file is corrupt.
    • Fixed the ability to start if there is an output port that has no value assigned and does not receive a value from a nested block (this behavior contradicts the requirement for Composite to always write all outputs).
  • DFBuilder: fixed GTDF/MaxParallel not being applied to the approximator used internally by the block.
  • File: fixed absolute file paths being stored in block configuration if the file is located into the project or sandbox directory. This bug is also fixed in all other blocks that use the file configuration component.
  • PythonScript:
    • Fixed a bug due to which the block was able to start without receiving all inputs.
    • Fixed incorrect behavior of the Add variables dialog in case when a test value is specified prior to selecting the type of variable.
  • Condition: fixed a bug which could break the block’s configuration if a variable is removed after renaming it.
  • Text
    • Fixed incorrect behavior of the Add variables dialog in case when a test value is specified prior to selecting the type of variable.
    • Fixed the block not showing an error message box when you click OK in the Edit script dialog and the script contains syntax errors.
  • UQ:
    • Fixed the inability to edit distribution parameters.
    • Fixed the appearance of the correlation matrix pane in the configuration dialog.
  • CSVParser: fixed a bug in file load due to which the block could report a non-existing file even if it actually exists.
  • Submatrix: fixed sending the signal to done more than once.
  • General:
    • Fixed the possibility to corrupt a project database by stopping the pSeven process while a workflow is running.
    • Fixed a bug in pSeven for Windows due to which NaN values input by user were converted to None, not allowing to set NaN as a port or variable value.
    • Fixed uninstaller not removing the pSeven Runner shortcut from system menus.
  • Block configuration:
    • Fixed incorrect name validation in blocks that support user-defined variables, which allowed to add two variables with the same name.
    • Fixed the ability to set block options that accept floating point values to None.
  • Workspace:
    • Fixed incorrect pSeven behavior after deleting the directory that is open on the Filesystem pane.
    • Fixed incorrect double-click handling in the pSeven example browser.
  • Edit: pSeven will now correctly resolve port name clashes when grouping multiple blocks with existing links that connect the grouped blocks to workflow root ports or other blocks outside of the group.
  • Run:
    • Fixed the ability to create a parameter preset with a name containing only whitespaces.
    • Fixed the ability to resize workflow annotations.
  • Analyze:
    • Fixed incorrect behavior of the “<last run>” record in the project database after renaming some record referenced by “<last run>”.
    • Fixed incorrect behavior of histogram plots when source data includes string values.
    • Fixed the inability to draw a 3D plot if there is a dataset with more than 3 data series in the plot configuration.
    • Fixed incorrect chart rendering and location of labels on stacked bar charts showing data with negative values.
    • Fixed incorrect report rendering when plot titles, labels, or names of data series contain HTML snippets.
    • Fixed incorrect rendering of parallel coordinates plots with negative values and non-linear axis scales.
    • Fixed a bug in parallel coordinates plot export which corrupted the exported image.
    • Fixed bugs in importing data from CSV files with complex headers.
    • Fixed an error when importing a CSV file that includes only the header line but no data.
    • Fixed the needless update of reports opened in background tabs when changes are made to the active report.
    • Varios minor corrections in the sample viewer.
  • Console: fixed internal block identifiers appearing in the log instead of block names after grouping some blocks to Composite.
  • GUI:
    • Fixed the broken help button in the Import workflow dialog.
    • Corrected tooltips in various menus and dialogs (view toolbar in Edit, data import and export, plot configuration and others).
    • Fixed ghost selection appearing in data editor dialogs for vector types if the vector length is 0 and a button to fill the vector with random values or zeroes is clicked.
    • Fixed resizing the Data series pane in Analyze.
    • Fixed the Delete and Change source… commands being available in the Data series pane context menu even if no data series are selected.
    • Fixed incorrect spacing between buttons on tab bars in Edit and Analyze when no tabs are open.

pSeven 6.0

Release date: 17 July 2015.

This release includes MACROS 6.0.

The following changelog lists all updates since pSeven 5.3, including the ones introduced in pSeven 6.0 Release Candidate 1 (the preview version released 26 June 2015). Changes that appeared in 6.0 Release Candidate 1 are noted (RC 1) in the changelog.

New Features

  • (RC 1) pSeven now provides native support for 64-bit Windows editions. The 64-bit version of pSeven for Windows is available from the DATADVANCE download page.
  • (RC 1) More examples are added to pSeven installation, including advanced examples of CAD tools integration and complex problem analysis. See section Examples for a usage guide.
  • (RC 1) New sample viewer tool in Analyze. You can use this tool to run a quick initial analysis of a data sample, collect descriptive statistics, and estimate correlations — see section Sample Viewer for details. Note that the sample viewer in fact replaces the table, descriptive statistics and scatter plot matrix viewers which are now available as its tabs.
  • (RC 1) Added Text, a new and improved block replacing TextGenerator and TextParser. These blocks are now deprecated. Old workflows containing these blocks will continue to work, so version compatibility is not affected.
  • All gradient-based methods in Optimizer are now capable to deal with cusp-like singularities in objective and constraint functions (codimension-one discontinuities of first derivatives). In particular, this can result in better performance when solving problems like structural optimization, where cusp singularities are common.
  • (RC 1) Added option presets in Optimizer. The presets quickly configure the block for solving a specific kind of problem: analytical problems, smooth or noisy problems, and so on.
  • (RC 1) The NX block has been released. This block features the following updates:
    • Linux version support.
    • The support for assemblies.
    • The support for NX part (.prt), IGES (.igs), Parasolid (.x_t), and STEP (.step) export formats.
  • (RC 1) Added the support for changing values of document variables with assigned units of measurement in SolidWorks.
  • Added data series handling functions to the Project database pane context menu. In particular, added an option to transpose a database record when forming data series.
  • (RC 1) Projects now have descriptions which are shown in Workspace after opening the project. The description is stored into the file in the project directory and can be edited in pSeven or in any text editor (descriptions use the Markdown syntax).
  • Workflow descriptions and annotations in Edit now support the Markdown syntax.
  • Added an edit button to various input fields which appears on hover and can be used instead of double-clicking the field.

Updates and Changes

  • Updated the Rotating Disk Optimization, Static Mixer Optimization, and Robust Optimization example projects:
    • All examples now use the new Text block in place of the deprecated TextGenerator and TextParser blocks.
    • Added reports filled with data for complex workflows that use external applications (SolidWorks, ANSYS). These reports are intended for users who do not have all required applications installed, so they can still study the results without running workflows. Note that most example reports contain no data but are configured in such a way that they show results after example workflows are run.
    • Added more details to example workflow annotations in Edit.
    • Added workflow descriptions shown in Run.
  • Added a reliability analysis example to the General Examples project. See the project description in Workspace for details.
  • Text improvements:
    • Operations added using quick selections are now tested automatically.
    • Enabled editing operations by double-clicking them on the Operations pane.
    • Write and insert operations in scripting mode now support Python tuples.
    • Added detailed error messages for operations with incorrect arguments.
    • Added a validation warning when a required input port is disconnected.
    • Added a validation error when an input file is not set.
    • Added more tooltips.
  • Sample viewer improvements:
    • Selecting a dimension pair on the Scatter matrix tab now automatically selects the corresponding sample columns on the Data tab as well as the related data series on the Data series pane. Note that it allows quickly creating a separate 2D plot of the selected dimension pair — just click the 2D plot button after selecting a scatter plot matrix cell.
    • Added detailed tooltips in the dependency measure selector on the Scatter matrix tab.
    • The viewer now provides trivial statistics for string data (total strings and the number of unique strings in the sample).
    • Increased the default scatter plot matrix budget from 300 to 1000 points.
    • Enabled navigating the sample viewer configuration window with the Tab key.
  • (RC 1) IVE now uses the new pSeven Core component, the Generic Tool for Sensitivity and Dependency Analysis (GTSDA).
  • Updated the FAST sensitivity analysis method in IVE so it provides more accurate estimates.
  • Parallel model training in ApproxBuilder is now disabled for small training samples by default due to computational inefficiency. See the GTApprox/MaxParallel option description for more details.
  • PythonScript now allows setting the default value of a RealScalar variable to None (no data), and also allows None elements in default values for RealVector and RealMatrix variables.
  • (RC 1) When the workflow is stopped by user, Program now does not wait until the task it started finishes.
  • If Program error handling behavior is set to “output defaults and signal”, error messages will now appear at the warning level in the console log.
  • The command mode is now default in the Program block configuration.
  • Better error handling in NX.
  • Added proper warnings on incorrect model file configuration in ApproxPlayer, DFPlayer, DRCompressor, and DRDecompressor.
  • More convenient project creation dialog.
  • Better appearance of project descriptions and the Welcome and Project description tabs in Workspace.
  • Default descriptions for new projects now contain a quick guide on editing the description and using the Markdown syntax.
  • The information pane in Edit now shows custom ports added to the workflow root block (workflow inputs and outputs).
  • Data editors for the Dict, List, vector and matrix types now support editing next-down cell on Enter and next-right cell on Tab.
  • Improvements of data filters in various viewers in Analyze:
    • Minimum and maximum can now be set to the same value, allowing to find all points containing exactly this value.
    • A filter can now be explicitly set to the [-Inf, +Inf] range. In particular, this allows filtering out table rows, which contain NaN values, since NaN is assumed to be out of any specified value range.
  • Checkboxes in various tables now can be enabled and disabled with the Space key.
  • pSeven error messages will no longer include information on the exception type which distracts users.
  • pSeven will now dim the background when showing configuration windows, message boxes and the like.
  • Various other minor improvements in user interface.
  • (RC 1) Updated some open source components to their recent versions, see section Open Source Components for details.


  • (RC 1) Many new tutorials and updated guides added to section Tutorials.
  • (RC 1) Updated section Examples now contains a guide on using the new example projects.
  • (RC 1) Updated section Introduction.
  • (RC 1) Added sections Workflow and Blocks containing more general information on editing pSeven workflows and configuring blocks.
  • Added the sample viewer description, see section Sample Viewer.
  • Added a guide on using the Keep globals option in PythonScript.
  • Updated section Open Source Components.

Compatibility Issues

  • (RC 1) Since 6.0 Release Candidate 1 pSeven supports changing values of document variables with assigned units of measurement in the SolidWorks block. A consequence of this update is that the block no longer converts output values to document measurement units for variables that were assigned their own units of measurement — thus block’s output in pSeven 6.0 Release Candidate 1 and above may be different from earlier versions.

    For example, consider a document where length units are meters and there is a variable specified in millimeters (x = 5mm). The block reads this value from document and outputs it to a port.

    • In earlier versions the output value is 0.005 (meters, value of variable converted to document units).
    • In pSeven 6.0 Release Candidate 1 and above the output values is 5 (millimeters, units of measurement are kept).


  • Fixed a bug in the 64-bit installer for Windows due to which the 64-bit pSeven version was installed to the Program Files (x86) folder.
  • The pSeven installer will now install pSeven in its own directory when selecting a non-default install location, instead of installing all package contents into the specified directory itself.
  • Fixed incorrect pSeven behavior on systems with Python-related environment variables set. pSeven will now ignore these environment variables because they interfere with the pSeven’s integrated Python interpreter.
  • Fixed a bug in pSeven for Windows which made it impossible to upgrade a project in versions 5.3 and up.
  • Fixed a bug in NX due to which in the Windows version of pSeven the block crashed if the path to Siemens NX DLL files was not specified in block options. Now the default path is searched if the option was not set.
  • (RC 1) Fixed a bug in Kompas3D due to which it sometimes did not close the KOMPAS-3D application process.
  • (RC 1) Fixed the ability to create variables named “do” and “done” in Creo and SolidWorks which leads to a conflict with reserved port names (input and output signal ports do and done).
  • Text bugfixes:
    • Fixed wrong line endings when generating output file on Windows.
    • Fixed incorrect block behavior when the output file is not set.
    • Fixed incorrect block behavior when the input file is a file of the temporary origin.
    • More minor fixes.
  • Re-enabled the Mutual Information technique in IVE.
  • Fixed a bug affecting Integration blocks due to which they could produce incorrect validation errors.
  • Fixed incorrect loading of the UQ block from previous pSeven versions due to which changing the analysis mode in block configuration caused a crash.
  • Fixed a bug in multidimensional CDF and PDF data output in UQ.
  • Fixed a crash in UQ which could occur upon switching to the Options tab.
  • (RC 1) Fixed pSeven not saving values from monitored ports if there is no link connected to the port and the value assigned to port is used. Port value will now be saved to project database as expected. A similar issue is fixed for workflow inputs: if the workflow was started with default input values, they were not saved.
  • (RC 1) Fixed incorrect order of values in data monitored from parallel ports of Composite blocks.
  • Fixed several problems in configuring input and output files for various blocks.
  • Fixed a bug in configuring files of the temporary origin which could result in incorrect file creation.
  • Fixed some integration example workflows not working in 32-bit Linux distributions.
  • Fixed reloading a workflow several times when it is imported as an Composite block.
  • Sample viewer bugfixes:
    • Fixed the viewer not updating properly when switching datasets (samples).
    • The viewer now correctly calculates descriptive statistics for samples containing NaN values. It also distinguishes between NaN and missing values and provides details for both on the Statistics tab.
    • Fixed incorrect viewer behavior when there is no data or not enough data for visualization — for example, when the sample contains one column only, the Approximation tab will display a placeholder instead of the approximation errors plot; earlier in this case the whole viewer became inactive and the only way to fix it was editing its configuration.
    • Fixed gradient marker colors on scatter plot matrix not being applied properly.
    • Fixed the inability to display text strings on the Data tab.
    • Fixed the inability to process data series of different length.
    • Fixed the inability to export the approximation errors plot to an image.
    • Fixed incorrect logging which could lead to unexpected behavior on the Approximation tab.
    • Other stability and interface fixes.
  • (RC 1) Fixed a memory leak on closing block configuration windows.
  • Fixed a bug in 3D plots due to which they crashed when rendering some specific data samples.
  • Fixed name conflict resolution when adding data to the Data series panel in Analyze.
  • Fixed a few incorrect labels and tooltips in Random configuration window.
  • Added a link to the NX help page to the Block library pane in Edit.
  • Fixed incorrect file modification time being shown in the Filesystem pane tooltips in Workspace.
  • Fixed incorrect tab placement on the tab bars in Edit, Run, and Analyze after resizing the pSeven window.
  • Corrected block and link tooltips in Edit.
  • Fixed error tooltips not disappearing when a block rename dialog in Edit is closed by pressing Esc.
  • Fixed overlapping toolbars in Run on small screens.
  • Fixed changes to the project description not being applied after editing it with the built-in editor.
  • Fixed project description always scrolling to the top after switching pSeven screens.
  • Fixed incorrect behavior of the project description editor if there is an empty link in the text.
  • Custom parameter description now correctly overrides the default ones when two or more workflow parameters are assigned the same alias.
  • Fixed a bug in the workflow run log panel due to which some last output strings were not shown in the log.
  • Fixed the clear log button not clearing the log sources list.
  • Fixed the console button not being highlighted when a fatal error message appears in the console log.
  • Fixed incorrect pSeven console log behavior when a workflow runs for a long time.
  • Fixed wrong ordering of dataset tabs in viewer configuration windows in Analyze.
  • Fixed incorrect rendering of stacked bar charts for certain data samples.
  • Fixed too big point value tooltips on parallel coordinates plots in Analyze.
  • The version update notice in Workspace now correctly reports the release type for Beta and Release Candidate versions.
  • Fixed the appearance of pSeven message boxes.
  • The pSeven help viewer now opens external hyperlinks in the system browser instead of trying to load the pages by itself.
  • Various other minor bugfixes in user interface.

pSeven 5.3

Release date: 8 June 2015.

This release includes MACROS 5.3.


  • Several improvements in Creo:
    • Added the support for using Creo in parallel Composite blocks.
    • Added part export support.
    • Added the support for part and assembly files containing a version number in file extension.
  • Added the support for the STL (.stl) export format in the CATIA block.
  • Program improvements:
    • Added an option to override the remote system’s locale in the script environment when using SSH.
    • The node count in SSH settings can now be set to 0, meaning there is no limit.
    • When configuring Program in the simplified command mode, the path to executable is automatically quoted if it contains whitespaces.
  • Optimizer now outputs an intermediate solution to a multi-objective problem if the workflow is stopped while the block is active. In particular, it helps when using surrogate based optimization, where Optimizer iterations can become time-consuming due to retraining of the internal approximation models.
  • Improved data import in Analyze:
    • Added the support for importing data with missing values (for example, empty fields in a CSV file).
    • Better performance for big files.
  • Similarly to data import, CSVParser now also supports data with missing values.
  • Added the capability to change the order of layers on 2D plots by reordering corresponding tabs in the plot configuration dialog.
  • Added point tooltips, selection range labels, and selected point highlight on parallel coordinates plots.
  • Improved the cascade and tile auto-layouting of data viewers in Analyze.
  • Added the capability to copy and paste annotations in Edit.
  • Console input is now disabled while running a workflow.
  • Various minor user interface improvements.


  • Fixed a crash on renaming an output variable in UQ.
  • Fixed a bug in Kompas3D which caused a crash when KOMPAS-3D produces an error if the Windows system locale for non-Unicode programs is not set to Russian.
  • Fixed Creo crashing sometimes when the block cannot launch the Creo application.
  • Fixed the inability to load a part or assembly file containing cyclic dependencies in Creo.
  • Fixed slow loading of iTA models in ApproxPlayer.
  • Fixed Composite not caching output values if an output port is not connected.
  • Fixed incorrect cache handling when a cached Composite is nested into a parallel Composite block.
  • Fixed Program substituting relative file paths instead of absolute if the input and output file names are set.
  • Fixed Program creating empty output files before writing output, which in some cases prevents the program it launches from writing these files, breaking the workflow.
  • Fixed incorrect pSeven behavior when opening many workflows or reports at once.
  • Fixed a bug in the new project dialog which allowed to create a new project inside an existing one under Windows.
  • Fixed incorrect behavior of the project open dialog after removing the path in the address bar.
  • Fixed a bug in saving workflows on close due to which a workflow sometimes was closed before it is actually saved.
  • Fixed the inability to uplink a port which name contains whitespaces.
  • Fixed the inability to import or export data in Analyze under Windows if the path to file contains Unicode characters.
  • Fixed incorrect display of constant values on parallel coordinates plots.
  • Fixed incorrect rendering of column charts with display limit less than the number of categories.
  • Fixed incorrect automatic determination of axis ranges on 2D plots with error bars enabled.
  • Fixed reports not saving the order of data viewer windows.
  • Fixed a bug that sometimes caused messages disappearing from the pSeven console.
  • Various minor user interface bugfixes.

pSeven 5.2

Release date: 23 April 2015.

This release includes MACROS 5.2.


  • Introduced new license feature, pSeven Runner. This is a separate launch mode intended to provide more flexible licenses. pSeven Runner does not require a full license and allows only to open existing projects and run saved workflows. This mode disables Edit and Analyze but does not limit other functions — you can edit workflow configuration, configure monitoring, and save workflow results, which then can be post-processed in the full version of pSeven.
  • Explicit block synchronization is no longer required in workflows with two or more SolidWorks or CATIA blocks. In particular, it allows using SolidWorks or CATIA in parallel Composite blocks. Previously the issue was that when these blocks started in parallel, the workflow crashed because SolidWorks and CATIA COM server configuration does not allow processing multiple documents simultaneously — thus it was required to control block execution order manually using their do and done ports. After update this workaround is no longer needed because pSeven now controls the execution order automatically.
  • pSeven now prevents system standby in Windows when a workflow is running.
  • To make project directory structure more consistent, pSeven now requires all workflows and reports to be located at the top level of the project directory. Subdirectories in project are allowed for resource files, block sandboxes and other data.
  • Added the support for problems with multidimensional output in UQ.
  • Added options controlling the behavior of randomized training algorithms in DFBuilder (see GTDF/Deterministic, GTDF/Seed).
  • Workflow descriptions are now shown in Run by default.
  • Tables in Analyze now allow you to set a specific number format for each column.
  • Minor improvements in data import and export in Analyze:
    • If no data (columns, data series) was selected to export, the export dialog now selects all available data by default.
    • Result settings in the import dialog can now be automatically synchronized with column selections on the Preview tab.
  • Various minor user interface improvements.
  • Updated some open source components to their recent versions, see section Open Source Components for details.


  • Fixed the inability to upgrade a project to current version if the path to project directory contains Unicode characters.
  • Fixed an error when deleting a block sandbox or re-creating it from a sandbox prototype which occurred when the sandbox directory is locked by antivirus.
  • Fixed non-deterministic Composite cache behavior in case there are cached points whose inputs differ by a value close to machine epsilon. The bug could lead to a divergence in optimization history when an optimization workflow is run for a second time using the data cached from the first run.
  • Fixed Creo requiring administrator rights to start.
  • Fixed Program ignoring script execution timeout when using SSH.
  • Fixed incorrect behavior of the TextGenerator and TextParser blocks when there is an output variable which is not read from file, and corresponding output port is not connected (blocks stopped instead of ignoring this issue).
  • Fixed a redundant warning on disconnected dimension input in DRCompressor (this input is optional).
  • Fixed incorrect behavior of the Random block in the deterministic mode.
  • Fixed a bug in pSeven for Windows due to which a file or folder rename (in Workspace) that changes only the character case (for example, “SomeFolder” to “somefolder”) resulted in deleting it instead of renaming.
  • Fixed the inability to run a Python script in Workspace if the file extension contains uppercase characters.
  • Fixed the inability to execute a Python script in batch mode (p7 --run in pSeven for Windows if the path to the script contains Unicode characters.
  • Bugfixes to the information panel in Edit:
    • The panel now shows all ports of the selected block, not only the connected ports.
    • Port and option lists are now sorted by name.
  • Fixed incorrect parsing of strings "y" and "n" as Boolean true and false values in the value editor for string types (StringScalar, StringVector, StringMatrix).
  • Fixed incorrect parsing of non-numeric values in the value editor for the RealScalar type.
  • Fixed monitor descriptions in Run disappearing after some changes to the workflow that do not affect monitoring settings.
  • Fixed the ability to change workflow parameter values and monitoring settings while the workflow is running, which could confuse the user.
  • Fixed incorrect names of workflow parameters in the project database.
  • Fixed a crash in Analyze when many reports with autorefreshing plots are open.
  • Fixed a bug which could cause incorrect rendering of histograms in Analyze.
  • Fixed a bug which could break the order of columns in tables in Analyze.
  • Fixed incorrect display of negative values on column charts in Analyze.
  • Fixed a bug in the built-in file open dialog due to which in Windows it interpreted paths as case-sensitive.
  • Fixed a bug that sometimes caused incorrect log message order in pSeven console.
  • Various minor user interface bugfixes.


pSeven 5.1

Release date: 16 March 2015.

This release includes MACROS 5.1.


  • Added the Gradient Boosted Regression Trees (GBRT) approximation technique in ApproxBuilder. See section Gradient Boosted Regression Trees for details.
  • Added an option for deterministic mode in UQ and Random which allows reproducing sequences of pseudorandom values generated by these blocks.
  • Changed some default monitoring rules for Optimizer ports:
    • Ports that output extended result data (infeasible outputs) are now monitored by default.
    • In robust optimization, ports that receive values of stochastic variables from random generator are no longer monitored by default.
    • Added a validation warning in case when values of variables, objectives or constraints are monitored in robust optimization (due to the big amount of data).
  • The block information panel is now available both in Edit and Run.
  • Added file encoding and line endings selection to text viewer in Workspace.


  • Fixed Program not restoring an SSH connection after it has been lost.
  • Fixed Program not closing remote processes created over SSH, when the workflow is forced to stop by user.
  • Fixed incorrect file output from Program working over SSH in case when error handling behavior is set to “output defaults and signal” and the script finishes without errors.
  • Workflow validation now properly takes into account the settings done in workflow configuration (parameter values and monitoring settings). Previously, only block configuration was analyzed.
  • Fixed the inability to open a Windows network folder in the file open dialog.
  • Fixed the workflow import dialog (see Workflow Import) where you could not select a workflow file.
  • Fixed incorrect encoding of CSV files exported from Analyze.
  • Fixed the inability to close a 3D plot viewer while the plot is still being loaded.


pSeven 5.0

Release date: 20 February 2015.

This release includes MACROS 5.0.


  • Improved template and sample file viewers in TextGenerator and TextParser. The viewers now show the search scope, results of operations and errors if they are found when previewing the result.

  • Reworked options controlling the behavior of certain randomized algorithms used in approximation (ApproxBuilder, DFBuilder):

    For more details, see the option descriptions.

Compatibility Issues

  • This release includes several changes in options controlling the behavior of certain randomized algorithms used in approximation (ApproxBuilder and DFBuilder blocks).

    Prior to pSeven 5.0 there were several options controlling the behavior of randomized algorithms in certain ApproxBuilder techniques, namely GTApprox/HDAInitialization, GTApprox/HDARandomSeed, and GTApprox/SGPSeedValue. To simplify the options structure, these options were removed and replaced with two common options: GTApprox/Deterministic that switches between the deterministic and non-deterministic modes and GTApprox/Seed that sets the fixed seed used in the deterministic mode (if it is on). Also, GTApprox/Seed does not have a special value to generate a random seed (as it was for GTApprox/HDARandomSeed and GTApprox/SGPSeedValue) — this function is now entirely controlled by GTApprox/Deterministic.

    This change can affect compatibility due to option name changes, but default technique behavior and available controls remain the same.

    By default, model training in previous versions was deterministic; this behavior is kept as default (GTApprox/Deterministic is on by default). To change HDA, HDAGP and SGP initialization, you can now do one of the following:

    • Keep GTApprox/Deterministic on but change GTApprox/Seed (new deterministic initialization).
    • Switch GTApprox/Deterministic off. In this case GTApprox/Seed is ignored. This is random initialization, the same as setting GTApprox/HDAInitialization to "Random" or using special GTApprox/HDARandomSeed and GTApprox/SGPSeedValue values in previous versions.

    Similar changes were done in internal validation (cross validation) options in ApproxBuilder and DFBuilder to keep their names and behavior in line with the described above:

    • GTApprox/IVRandomSeed is renamed to GTApprox/IVSeed and simply sets the seed used if GTApprox/IVDeterministic is on. Default behavior is deterministic, the same as in previous versions.
    • GTDF/IVRandomSeed is renamed to GTDF/IVSeed and sets the seed used if GTDF/IVDeterministic is on. Default behavior is deterministic, the same as in previous versions.

    Also, to help reproduce the results obtained in a non-deterministic mode, the seeds generated for training and cross validation are now added to model info.

pSeven 5.0 Release Candidate 3

Release date: 13 February 2015.

This release includes MACROS 5.0 Release Candidate 3.


  • ApproxPlayer can now automatically export approximation models to C code, m-file and MEX-file.
  • C source files (.c), m-files (.m) and MEX-files (.mex) are now recognized by pSeven and can be opened in Workspace for viewing.
  • Added ability to validate a model on a reference data sample in ApproxPlayer and DFPlayer.
  • Run improvements:
    • Merged parameters and monitoring settings into a single Configuration tab.
    • Added Run log tab displaying run-time information separately for each workflow.
  • Added automatic design-time validation of block configuration in Edit and plot configuration in Analyze. Validation warnings are shown in the new issues console separated from the pSeven console and workflow run logs.
  • Changed the view of the pSeven console. Due to the addition of the issue console and workflow run logs, the main console is now intended mainly for debugging.
  • Analyze improvements:
    • Added data filtering for tables and plots allowing to apply value-based filters to select certain subsets of source data to be displayed.
    • Added point distribution histograms and correlation information in the scatter plot matrix viewer.
    • Added a quick access button to view a selected pair from a scatter plot matrix as a separate plot.
    • Minor redesign of plot configuration dialogs.
  • Since this version, created projects, workflows and reports will store exact version information. When available, version number is now shown in a tooltip when you select a file in Workspace (files from versions prior to pSeven 5.0 will be only labeled as outdated).
  • If you try to open a locked project, pSeven is now able to check if this project is actually open in another pSeven instance and if not, the lock is removed automatically. This cancels the need to remove the lock file manually in case when a project was left locked due to a previous crash.
  • Minor usability improvements in value editors for BoolScalar, vector and Dict types.


pSeven 5.0 Release Candidate 2

Release date: 4 February 2015.

This release includes MACROS 5.0 Release Candidate 2.


  • Added NX, the block for Siemens NX integration. This block is in development and is currently found in beta blocks in the Block Library.
  • Added simplified configuration mode in Program. In the new new command mode, you can select a program to run, specify its arguments and add input and output files without actually writing a script.
  • Improved workflow validation in cases when some block ports have been assigned default values or were added as parameters to workflow configuration.
  • Reduced the Linux installer size.


  • Fixed Program sometimes not capturing the standard error stream (stderr) properly.
  • Fixed unexpected data type compatibility issues when UQ distribution parameters are set from workflow configuration.
  • Fixed an “already open” error in case when a project is opened by double-clicking a workflow or report that was not closed manually when closing the project in the previous session.
  • Fixed errors in resolving data types for uplinked ports.

pSeven 5.0 Release Candidate 1

Release date: 29 December 2014.

This release includes MACROS 5.0 Release Candidate 1.


  • Added the block information panel in Edit. This panel can be used to quickly review any block configuration: when you select a block, it shows which block settings were changed from defaults, including values assigned to ports, changed options, and various block-specific settings.
  • Improved the handling of parameter names containing special characters in Creo, Kompas3D, and SolidWorks to prevent unwanted behavior in case of complex parameter names.
  • Added an option to control error handling behavior to all Exploration, Modeling and Integration blocks. When a block supporting this option encounters an error during execution, it can either stop the workflow (default), stop itself and send a failure signal to the done port, or stop itself and output a signal to done as well as the values assigned to other output ports. In the two latter cases, the failure signal can be handled, for example, by Condition, allowing to continue executing the workflow in case of block errors.
  • Redesigned block configuration dialogs with left-side tabs.

pSeven 4.3

Release date: 24 December 2014.

This release includes MACROS 4.3.


  • Fixed incorrect behavior of UQ when analysis starting point is set.
  • Fixed incorrect calculation of correlation coefficients in the descriptive statistics tool in case when data samples contain NaN values.

pSeven 4.2 Service Pack 1

Release date: 16 December 2014.

This release includes MACROS 4.2 Service Pack 1.


  • Fixed Optimizer crashing if the problem is stochastic and unconstrained, and monitoring is enabled for the result ports that output point feasibility measures.

pSeven 4.2

Release date: 10 December 2014.

This release includes MACROS 4.2.


  • Optimizer can now output an extended result data set. This set includes additional optimal solutions that violate problem constraints up to a certain (controllable) extent. To get an extended result, enable infeasible outputs in Optimizer configuration.
  • Added the ability to include constraint-wise point feasibility measures in Optimizer result. This metric is a normalized distance to the closest (lower or upper) constraint bound calculated for each point in the solution per each constraint. It can be useful in the following ways:
    • For feasible points (optimal points which strictly satisfy constraints), it can show which constraints are “active” for a point, that is, a constraint is close to bound.
    • For infeasible points (optimal points from the extended data set, see above), it can show which constraints are violated and how bad the violation is, compared to the value of the respective constraint bound.
  • When a variable is added in UQ, the block now automatically creates a new option that sets distribution parameters of this variable. Such options can be added to workflow parameters, as a result allowing to configure UQ variable distributions in Run.
  • UQ and Distribution now monitor more ports by default.
  • Updated Tensor Approximation (TA) and incomplete Tensor Approximation (iTA) techniques in ApproxBuilder improve TA and iTA model quality in certain cases with noisy training data.
  • Added exact fit support to the incomplete Tensor Approximation (iTA) technique: ApproxBuilder can now train iTA models exactly fitting training data points.
  • When you rename (copy, move) a workflow or report in Workspace, pSeven now automatically adds the correct file extension. This also fixes a bug due to which a report database could be lost if the new report name is specified without extension.
  • Added a limit on the number of presets (128) that can be created when importing a workflow into Composite. For details, see section Workflow Import.


  • Fixed a bug with file permissions in Program working over SSH and configured to use a sandbox prototype. The block did not preserve permissions when files are copied from a sandbox prototype to the remote sandbox, as a result it was unable to execute scripts on a remote host. This bug did not affect local to remote sandbox synchronization — that is, to encounter it, a sandbox prototype must be set in addition to local and remote sandbox directories (see section Sandbox for details).
  • Fixed a bug in the Tensor Approximation (TA) technique in ApproxBuilder due to which TA models output nonsensical gradient values for discrete variables. All partial derivatives with respect to discrete variables are now NaN.
  • Fixed Creo not shutting down Creo processes properly if there was an error when loading a document.
  • Fixed the incorrect display of legends on 2D plots when they are exported to images (the legend did not show point markers).

pSeven 4.1

Release date: 27 November 2014.

This release includes MACROS 4.1.


  • Increased startup performance of all blocks. pSeven GUI also starts faster and loads CPU less on startup.
  • Added mixed integer linear programming support in Optimizer. Single-objective problems with a mix of integer and continuous variables can now be solved without using the Surrogate Based Optimization (SBO) mode, provided that all problem functions are linear (linearity type has to be set when adding objectives and constraints in Optimizer). Note that this feature requires analytical (user-defined) gradients for all objective and constraint functions.
  • Overall increase in Optimizer performance as a result of significant improvements of internal algorithms — in particular, the methods to solve saddle point systems and quadratic problems.
  • Added point batch generation support in UQ. This mode allows you to take advantage of parallel processing in uncertainty quantification workfows by enabling batch generation in UQ and wrapping the analyzed model into an Composite block with parallelization ratio equal to the batch size.
  • Added an option to set the Creo environment file path in Creo. This allows using Creo when PRO_COMM_MSG_EXE is not set and selecting another installed version of PTC Creo without changing the system environment variable.
  • Removed redundant ports for blackbox bounds in DoE (lower_bounds and upper_bounds) which appeared when the block is switched to the adaptive DoE mode.
  • Added the ability to duplicate plots and tables in Analyze — creates an exact copy of an existing plot under a new name. The function is available from the plot and table viewer’s context menu.
  • Plot legends in Analyze now show both marker and line colors.
  • Improved performance of the pSeven console logger.


  • Fixed Creo being unable to build the document tree when dimensions are not specified in the loaded model.
  • Fixed PythonScript crashing after some variables are renamed in such a way that their names are swapped (for example, “x” to “y”, “y” to “x”).
  • Fixed an incorrect behavior of the file paste function in Workspace in case when the file is renamed after copy or cut.
  • Fixed incorrect error messages when attempting to load a document into Kompas3D or Creo block on a machine where KOMPAS-3D or PTC Creo is not installed.

pSeven 4.0

Release date: 5 November 2014.

This release includes MACROS 4.0.


  • Changed default configuration of the local license server (used when pSeven runs in the standalone mode) to avoid port conflicts with other software using the FlexNet licensing system. This change can affect version compatibility – see Compatibility Issues section below for details.
  • UQ block has been released.
  • Added scatter matrix plot tool in Analyze.
  • Added model validation data output to ApproxPlayer (the iv_info port).
  • The Program block now always creates output files, even if the script did not write the actual result file (the block outputs an empty file in this case). This change may affect compatibility – see Compatibility Issues section below for details.
  • Contents of Filesystem and Project panes in Workspace are now automatically refreshed when changes are detected.
  • Added new column chart viewer in Analyze which replaces old bar and column chart viewers. The new version adds more settings and allows fast switching from chart to histogram representation.
  • The descriptive statistics viewer is removed from tables and plots and added as a separate viewer in Analyze.

Compatibility Issues

  • pSeven 4.0 changes the listening port of the local license server that provides licenses to pSeven running in the standalone mode (that is, using a local license file). Previous versions used default FlexNet Publisher port 27000, which in many cases lead to port conflicts with other software using the FlexNet licensing system. pSeven 4.0 uses port 27042 to avoid this typical conflict. However, if a license file from an older version is found on the host where pSeven is installed, the local license server may cease to start because of different port numbers in the old and new license files.

    This issue is identified by the following message in pSeven logs:

    ERROR: license files have conflicting port numbers
    Please edit the license files so the port numbers on the SERVER line are identical.


    pSeven logs are found under the pSeven settings directory (%USERPROFILE%\.p7\ in Windows, ~/.p7/ in Linux). In Windows, you can also use a menu shortcut to open this directory: Start ‣ All Programs ‣ pSeven ‣ pSeven Logs.

    The recommended solution is to uninstall older versions of pSeven and switch to using pSeven 4.0. Note that the pSeven uninstaller does not remove user files and settings, so old license files can be left in the pSeven settings directory. These files have to be removed manually.

    pSeven settings directory is %USERPROFILE%\.p7\ in Windows and ~/.p7/ in Linux. You can also open this directory with a system menu shortcut (for example, in Windows: Start ‣ pSeven ‣ License Management ‣ Open License File Location).

    If needed, it is possible to keep both pSeven 4.0 and an older version installed and functional, but solution depends on how pSeven was installed, the number of users, and other details. You can contact for help with this configuration.

  • This release changes behavior of output file ports of the Program block. Previously, if the block’s script failed and did not generate some file that would be output to a port, this port did not output anything. Since pSeven 4.0, in this case an empty file is output instead.

    Old workflows assuming there is no output from Program in case of script failure are likely to behave not as expected when run in 4.0. It is recommended to review settings of blocks which parse Program output files. If you need to ignore empty files from Program, you can also use a Condition block with a condition based on script exit status.

  • pSeven 4.0 introduced certain changes in generating parameter names in the SolidWorks and Kompas3D blocks when document parameters are added to block parameters (mapped to ports). When upgrading a project from version 3.4 or older to 4.0, this can result in block configuration being broken if in the original document there is a parameter which name contains one of the following characters: < > : " \ | ? *. In 4.0 these characters are prohibited in document parameter names and will be replaced; as a result, the associations between block parameters and document parameters will be lost, and the block will not be able to start.

    This situation is fixed in pSeven 4.1 and above by allowing all characters in block parameter names, so if a project is upgraded from 3.4 or older to 4.1 or newer, it will continue to work normally. However, if the project was first upgraded from 3.4 or older to 4.0, and then from 4.0 to 4.1 or newer, the broken block configuration is not fixed automatically, and the project will continue to be unusable.

    Due to this, it is recommended to skip version 4.0 if you require the support for parameter names such as “Part/Subpart” (or containing other characters from < > : " \ | ? *) and upgrade to 4.1 directly.

    If you are upgrading to 4.1 or newer, it is recommended to restore a backup of the project from version 3.4 and upgrade it to the current version, skipping version 4.0. If this is not possible, you can manually edit block parameters after upgrading the project to the current version to make block parameter names the same as document parameter names. Another option is to remove all block parameters and re-add them, allowing pSeven to generate correct block parameter names automatically.


  • Fixed a bug due to which on some systems pSeven found a standalone license file only if it is placed under the user directory and could not find a license in the installation directory (see Standalone License Registration for the details of this setup).
  • Fixed a bug which could cause pSeven instability when many blocks generate errors at once.
  • Fixed a module import error in Creo under Linux.
  • Fixed Composite block cache being corrupted when the workflow is stopped or terminated by user.
  • Fixed incorrect column ordering in table configuration dialog after rearranging columns in a table viewer in Analyze.
  • Fixed excessive memory consumption when zooming a 2D plot that contains points with coordinates equal to maximum representable floating point number.
  • Fixed uncommon behavior of file copy and paste functions in some cases.


pSeven 4.0 Release Candidate 1

Release date: 15 October 2014.

This release includes MACROS 4.0 Release Candidate 1.


  • If an outdated workflow is discovered while opening a project, pSeven now automatically upgrades it to the current version.
  • Added a button to toggle workflow annotations to the view toolbar in Edit and Run.


  • Fixed incorrect order of elements in monitoring data collected from output ports of Composite block with enabled parallel processing. In project database, output order did not match input order. Nevertheless, workflows with parallel Composite blocks worked correctly, as the run-time was not affected by this bug.
  • Fixed pSeven freezing when you add a variable with inf as the default port value in PythonScript or add an inf constant in Const.
  • Fixed 3D plots not drawing trajectories when point markers are disabled.
  • Fixed incorrect rendering of a plot with logarithmic axis scales after exporting this plot to an image.

pSeven 4.0 Beta 1

Release date: 1 October 2014.

This release includes MACROS 4.0 Beta 1.


  • New block Creo that enables integration of PTC Creo parts and assemblies into pSeven workflows.

  • Major improvements in the Surrogate Based Optimization (SBO) method in Optimizer:

    • The method now uses a new family of internal surrogate models with greatly decreased training time and complexity which are specifically tuned for SBO.
    • Internal SBO algorithms are now multi-scale and multi-resolution capable, meaning that functions with intricate landscape could be modelled easily.

    Note that the updated SBO algorithms are slightly more demanding with respect to required budget: the total time to solve usually decreases due to the increased algorithm efficiency, but new algorithms require more function evaluations.

  • Monitoring improvements:

    • Added last run shortcut in project database. The shortcut always points to the monitoring data obtained from the latest workflow run. This allows quickly updating tables and plots that are configured to use the last run record as a data source.
    • The values of workflow inputs and outputs are now stored separately from other monitored data.
    • Run records now also store parameter values that were set on the Parameters tab in Run and run statistics such as start and finish timestamps, the number of times each block started, and other.
    • Run name can be set on the Monitoring tab in Run. Run name is the name of the project database record that will store the data monitored in this run.
  • Data series in Analyze can now be created from individual entries in database records, not only from entire records.

  • Added histogram plot type in Analyze.

  • Added descriptive statistics to table and plot viewers in Analyze — provides univariate analysis and correlation information.

  • Added option name filtering in block configuration dialogs, Options tab.

  • When importing a workflow to an Composite block, input and output port names for this block are now copied from aliases set for workflow inputs and outputs. This makes new port names the same as those seen in Run for the imported workflow (previously, import copied the original port names which could be ambiguous).


  • Fixed a crash on creating a 3D plot that uses a data source containing only NaN or Inf values.
  • Fixed incorrect behavior of CSVParser configuration dialog after an attempt to load a non-existing file.
  • Fixed blackbox-based adaptive DoE in DoE always including all points of the initial sample into the final result, even if some of them violate specified generation bounds. Now the result will include only those initial points that do not violate the bounds.
  • Fixed the inability to rename or remove a project subdirectory in the Project pane, if the same directory is open in the Filesystem pane in Workspace.
  • Fixed incorrect information on block ports displayed in tooltips for the currently edited Composite block.
  • Fixed incorrect behavior when opening workflows or reports copied from a project created in an older pSeven version to a project created in the current version.
  • Fixed incorrect behavior when importing a workflow created in an older pSeven version into an Composite block.


pSeven 3.4

Release date: 18 September 2014.

This release includes MACROS 3.4.


  • Merged the general block properties dialog with the configuration dialog. As a result, the following features are now available as new tabs in all block configuration dialogs:
    • General options editor (Options tab), which also allows selecting any option as a workflow parameter. You can review the list of parameters in the workflow configuration dialog (Parameters tab) and set parameter values in Run.
    • Input and output port editor (Ports tab) where you can change default values for ports, select ports for monitoring, and select any port as a workflow parameter. In Composite configuration dialog, the port editor also allows adding and removing ports.
    • Environment variables editor (Environment tab) where you can set environment variables for the block’s process.
    • Block’s working directory settings (Sandbox tab) — see section Sandbox for details.
  • Improved workflow run statistics window now shows detailed information on missing inputs and other problems encountered when running a workflow.
  • Added block and port name filtering to workflow configuration dialogs for adding parameters and monitored ports.
  • Added workflow tree viewer (shows block hierarchy) to the breadcrumbs in Edit.


  • Fixed pSeven converting absolute file path to relative for files of project origin located outside the project directory, in case when the file is located one level above the current project directory and filename (without extension) is the same as the project directory name.
  • Fixed File not creating a new empty file on start when a non-existent path is specified for a file of project or sandbox origin.


pSeven 3.3

Release date: 29 August 2014.

This release includes MACROS 3.3.


  • Improved workflow import and export to/from Composite block can now keep workflow input values, configuration and monitoring settings. The import and export functions have been moved from Composite block configuration window to the workflow edit toolbar, see section Workflow as a Block for more details.
  • Composite block now allows creating multiple parallel input ports, previous limitation (only one parallel input) is removed.
  • Parallel coordinates plots are now available in Analyze.
  • Updated uplink dialog to sort all input ports before outputs for each block in the list.


  • Fixed pSeven crashing when network connection to license server is lost. pSeven will now pause and try to restore connection, displaying an information dialog which also allows saving current changes and exit without waiting.
  • Better error handling for reports: opening a corrupted report should no longer freeze pSeven.
  • Fixed pSeven crashing when you run a workflow with nested Composite blocks which have parallel processing turned on.
  • Fixed the inability to copy workflow output values from the Outputs pane in Run to clipboard.
  • Fixed incorrect conversion of default values of PythonScript variables which crashed pSeven when changing the type of a variable with default already set.
  • Fixed a bug in path conversion for files of project origin which could apply unexpected transformations to absolute paths when a project created under Windows is edited and changed under Linux.


pSeven 3.2

Release date: 12 August 2014.

This release includes MACROS 3.2.


  • Added the ability to test SSH connection in Program block configuration.


  • Fixed a bug in workflow configuration due to which some blocks could ignore settings for their ports that are added as parameters to workflow configuration.
  • Fixed Program crashing when a file that should be output to port is not found after executing the script. The block will now correctly report an error to its status port, allowing the workflow to continue.
  • Fixed Program not working in pSeven for Windows if the path to executable contains special symbols.
  • Fixed a bug in PythonScript validation which crashed pSeven when different input and output defaults are specified for a block variable.
  • Fixed IntScalar inputs (integer model parameters) not working in the SolidWorks block.
  • Fixed the inability to hide console after cancelling a project upgrade.

pSeven 3.1

Release date: 18 July 2014.

This release includes MACROS 3.1.


  • Composite can now store its input-output cache in memory without writing to disk. Note that this memory cache is not persistent and clears when the workflow stops.
  • pSeven now automatically monitors important ports of the Map, Program, ApproxBuilder, DFBuilder, DFPlayer, DoE, DRBuilder, IVE, and Optimizer blocks.
  • Added confidence interval output to the UQ block.
  • Added an option to linearly extrapolate lines on a 2D plot.
  • Improved uplink dialog.
  • Various user interface fixes and improvements.


  • Added pre-run cache file validation in Composite to warn if the cache is malformed, or there is a port name or type mismatch.
  • Fixed SolidWorks sometimes not shutting down SolidWorks processes properly which led to a workflow freeze.
  • Fixed incorrect name checks for parameter and port aliases in workflow configuration.
  • Fixed a synchronization error which occurred when a workflow input or output is added as a parameter in workflow configuration.
  • Fixed a synchronization error in case when a block with monitored ports is removed from a workflow.
  • Fixed incorrect project path detection which could cause problems when using files of project origin.
  • Fixed positioning of copied and grouped blocks.
  • Fixed block and link tooltips in Edit not disappearing sometimes when mouse cursor is moved away.
  • Fixed zoom in Edit being reset when opening a block configuration window.
  • Fixed pSeven not remembering the position in workflow hierarchy when switching between open workflows.

pSeven 3.0

Release date: 27 June 2014.

This release includes MACROS 3.0.


  • Redesigned Workspace screen now provides quick access to project management functions (such as creating and adding projects, workflows, and reports) and various pSeven guides and examples.
  • Added file tooltips in Workspace which can contain extended information. For example, if a workflow selected in Filesystem or Project pane has a description, it is shown in the tooltip.
  • New workflow configuration tool available in Edit and Run offers many advanced options and allows creating workflows that are fully usable from the Run screen only. Simplify your workflow for its users, hiding unwanted complexity and allowing them to focus on the real task:
    • Select any port or block option as a parameter that is seen and edited in Run, for quick access to key settings.
    • Use parameter ports to specify block input values: no more need to clutter the workflow with Const blocks and to edit their configuration in order to change inputs.
    • Set parameter values when editing the workflow, and save them to configuration presets. You can create multiple presets to simplify workflow configuration even further or provide user with some starting points.
    • Add aliases and descriptions to parameters you create to help users understand their task-specific meaning without studying how they are processed in the workflow.
    • Parameter aliases are names shown in Run instead of actual port or option names. Multiple parameters can be given the same alias, so Run user only needs to change one value if there are many ports or options that should receive the same value.
    • Select ports that should allow monitoring and add them to workflow configuration. Only these ports are now shown in Run instead of the hardly comprehensible list of all ports in the workflow found in previous versions.
    • Just like for parameters, add aliases and descriptions to monitored ports to explain their meaning.
    • Finally, workflow configuration may contain a workflow description visible in Workspace tooltips.
  • Redesigned Run screen to follow the ideas behind the new workflow configuration tool:
    • New screen layout pays more attention to workflow inputs and outputs, parameters and monitoring settings. Workflow structure is shown as a mini preview instead of taking up the majority of screen.
    • In Run, users work with the simplified configuration interface and can switch presets, change parameter values and (de)select ports for monitoring if they decide to do so.
    • Workflow output values are now shown in Run when execution finishes, so main results can be seen without switching to Analyze.
    • Workflow inputs and outputs are automatically monitored: all values are saved to project database and can be studied in Analyze later.
  • Improved uplink: now if the uplinked port changes type (the block is reconfigured), the corresponding top-level port updates its type automatically, so the uplinks are not removed when editing blocks nested into Composite. This feature works only if the top-level port connects exclusively to the uplinked port, which is the usual case.
  • Improved value editor:
    • Allows in-line editing, useful for simple data and quickly copying values in their string representation.
    • Improved data type handling: when you change the type, values are not lost if they can be converted to the new type, and a warning is raised if lossless conversion is not possible.
    • Vector and matrix editors: fill vectors and matrices with zeros or specified value, or generate random.
    • Matrix editor: generate an identity matrix; transpose current matrix.
    • List and Dict editors: remove selected elements or keys, allows multiselection.
  • Redesigned plot configuration dialogs: plots of the same type (for example, all 2D plots or all charts) now have a configuration dialog common for this type. New dialogs also feature better settings layout and overall higher usability.
  • Fixed axis ranges can now be set in general plot configuration. No more need to manually zoom the plot if you want it to show a selected area only.
  • Improved 3D plots:
    • Data labels can be added to points.
    • The plot can show stems projecting points on the XY plane.
    • Added colorbars showing the color scale for Z axis.
    • Added ability to map gradient colors to point index in plot source data — useful for plotting trajectories and other ordered point arrays.
    • Fixed axis ranges can be set in plot configuration.
  • Added 3D surface plots supporting:
    • Various surface reconstruction methods (2D grid, 2D triangulation, 3D triangulation).
    • Gradient surface fill and colorbars.
    • Surface transparency.
    • Fixed axis ranges.
  • New features of Composite blocks:
    • The block can now import a workflow from file, making it a component of the current workflow (that is, imported workflow becomes the content of Composite block).
    • Automatic parallel processing: you can switch the block to parallel mode and it will automatically process an input list in parallel. The allowed number of parallel processes can be also set in block configuration.
    • Data cache support: the block can cache input and output values to a file specified and use it to speed up processing. If next input value is found in cache, the corresponding output value is read from cache too, instead of recalculating it. The cache file is a simple CSV that can be easily used outside pSeven.
    • Changed the rule for Composite block input ports: now all inputs are required to start the block. This may cause compatibility issues with workflows exploiting the old behavior of Composite blocks. See Compatibility Issues section below for details.
  • Kompas3D (previously added to beta blocks as Kompas) has been released. This block features the following updates:
    • New document tree viewer supporting hierarchy and complex trees.
    • Port names can now be automatically generated based on parameter names. In particular, this allows batch mapping of parameters to ports.
    • Full support for assemblies and variables with assigned units of measurement.
    • The support for ACIS (.sat), IGES (.igs), Parasolid (.x_t), STEP (.step), STL (.stl), and VRML (.wrl) export formats.
    • Many bugfixes in document exploration and better error handling.
  • Surrogate based optimization (SBO) method of Optimizer now supports mixed-integer problems. To use this feature, at least one of the problem objective or constraint functions should be configured as expensive (set evaluation cost type to “Expensive” when adding objective or constraint in block configuration).
  • Optimizer is now able to analyze behavior of linear and quadratic objectives and constraints and restore analytical functions so they are evaluated internally by the block in order to speed up solving.
  • ApproxBuilder features new Tensor Gaussian Processes (TGP) approximation technique:
    • TGP is a further development of the methods first introduced in the Tensor Approximation (TA) technique.
    • TGP modifies the Gaussian Processes (GP) algorithm for big data sets, provided they were obtained using a Cartesian product DoE.
    • A distinctive feature of the TGP technique is that it provides model accuracy evaluation support not available in the original TA or incomplete Tensor Approximation (iTA) techniques.
  • ApproxBuilder now supports model internal validation for the Mixture of Approximators (MoA) technique.
  • Updated Const block simplifies editing constants: values can be changed without bringing up a value editor, and data type for a constant is set automatically based on its current value.

Compatibility Issues

  • This release changes the rule of processing Composite block’s input ports: now all inputs are required to start the block. Earlier, the Composite block could start as soon as there was an input for at least one of the nested blocks. Old workflows exploiting this behavior will possibly freeze if run in 3.0 Final, because of Composite blocks waiting for all inputs, and have to be redesigned to avoid this.

    This change was required by the new feature of loading an existing workflow into a Composite block. Old behavior is in fact against the idea of the Composite block as an integral workflow component, thus the change makes the workflow model more consistent and will benefit in creating re-usable workflow components as custom Composite blocks.


  • Fixed a bug in compatibility with version 1.7.x-1.8.x workflows containing Modeling blocks which blocked upgrading them to the current version.
  • Fixed Program locking local files after uploading them through SSH.
  • Fixed Program substitutions not working when they appear in the middle of a command line and are not surrounded with whitespaces. Recommended syntax is now @{substitution_name}, for example: program.exe -in file_@{num}_in.dat -out file_@{num}_out.dat.
  • Fixed TextParser not starting if the sample file is not found or set to none, despite this file is only needed to configure the block but not at run-time.
  • Fixed incorrect behavior of Kompas3D block in case when it loads a document that is already opened in KOMPAS-3D application.


  • Added detailed TextGenerator block documentation.

pSeven 3.0 Release Candidate 2

Release date: 10 June 2014.

This release includes MACROS Release Candidate 2.


  • pSeven now integrates the OpenTURNS methodology: new Distribution and UQ blocks perform an uncertainty quantification study in a workflow. Note that both blocks are currently in beta state.
  • Added file and folder tooltips in Workspace.


  • Fixed Program always returning exit code 0 instead of the actual value when used with SSH.
  • Fixed PythonScript validation routine which crashed if the script contained a string with a null byte.
  • Fixed report file (.p7rep) rename or move in Workspace breaking its connection with the report database (.p7rdb). Now when you rename, copy or move reports in pSeven, database files are automatically renamed, copied or moved along with report files. Also, .p7rdb files are now hidden in Workspace to avoid confusion.
  • Fixed plot and table behavior in Analyze in case when some data sources are missing (for example, a source data series was deleted from database after configuring a plot).
  • Fixed incorrect colors and legend names in chart plots in case when the plot contains a hidden chart.

pSeven 3.0 Release Candidate 1

Release date: 26 May 2014.

This release includes MACROS 3.0 Release Candidate 1.


  • Redesigned Workspace screen with more useful shortcuts.
  • Added block and link tooltips in Edit. Block tooltips show information on input and output ports. Link tooltips provide connection details (linked ports and data flow direction).
  • Workflows can now contain sticky notes visible in Edit and Run. To add a note, double-click the workflow edit area or use the New annotation button on the edit toolbar.
  • More convenient data editor dialogs for specifying workflow inputs and variable values.
  • The std block package now includes all blocks that are installed with pSeven by default. Package names adapters and macros are currently unused.
  • pSeven licensing system was changed and now uses floating type licenses only. pSeven continues to support standalone installations by using a local license server which is automatically installed and started with pSeven. See the updated License Setup chapter for details.


  • Added an error message when opening a project that is read-only or is locked by another instance of pSeven.
  • Fixed a run-time bug which made a workflow freeze, instead of terminating properly, if any block crashes while there is an unprocessed output from another block.

pSeven 3.0 Beta 2

Release date: 28 April 2014.

This release includes MACROS 3.0 Beta 2.


  • Fixed pSeven crashing sometimes after clicking Wait or Terminate in the workflow stop dialog.

pSeven 3.0 Beta 1

Release date: 12 March 2014.

This release includes MACROS 3.0 Beta 1.


  • New block Condition which effectively replaces the If, Switch, and Toggle blocks. The latter are now deprecated and are no longer available from the block library. Old workflows containing these blocks will continue to work, so version compatibility is not affected.
  • Added shortcuts to the default workspace directory, user directory and desktop to file open and save dialogs.
  • Redesigned common pSeven dialogs to make their behavior and controls consistent.
  • Removed the Actions pane from Workspace. Its functions are now accessible from the context menus in the Filesystem and Project panes.


  • Fixed pSeven not switching to a workflow that is already open in Edit when you double-click it in Workspace.
  • Added error messages for invalid file operations in Workspace.
  • Fixed case-sensitive path interpretation in the recent projects list under Windows.

pSeven 2.4

Release date: 15 May 2014.

This release includes MACROS 2.4.


  • Fixed excessive memory consumption when pSeven runs for a long time in graphical mode.
  • Fixed a version compatibility issue in SolidWorks block.
  • Fixed a bug in creating block working directory in Windows due to which a workflow could sometimes fail to start.

pSeven 2.3

Release date: 23 April 2014.

This is a maintenance release which does not contain any notable changes. Includes MACROS 2.3.

pSeven 2.2

Release date: 16 April 2014.

This release includes MACROS 2.2.


  • Many improvements in SolidWorks blocks:
    • Full support of assemblies with several components.
    • Improved SolidWorks document introspection: values, types and units for all parameters are now shown in design-time.
    • New document tree viewer supporting hierarchy and complex trees.
    • Added the support of ZIP archives (pack&go) for document and export files.
    • Added the support of different port types, not limited to RealScalar.
    • Port names can now be automatically generated based on parameter names. In particular, this allows batch mapping of parameters to ports.
    • Full support of file component, including files of port origin, working with relative paths and so on.
    • Better support for global variables allows mapping them to output ports.
    • Improved error handling.
  • You can now set prototype for block sandboxes — a directory used to initialize the sandbox when it is created (see section Sandbox for details).
  • Increased performance of Program blocks thanks to SSH sessions reuse.
  • For security reasons, SSH keys used by Program blocks are no longer saved with the workflow.


  • Fixed an important bug in Optimizer related to multi-objective Surrogate Based Optimization. Incorrect processing of constraints at anchor search stage in this mode could lead to selecting infeasible solutions as anchor points, potentially causing problems in solving.
  • Fixed pSeven checking out an extra license from floating license server when starting a workflow.
  • Fixed a bug in generating identifiers for files of temporary origin, due to which multiple blocks could use the same temporary file if they were created using copy and paste in Edit.

pSeven 2.1

Release date: 24 February 2014.

This release includes MACROS 2.1.


  • Optimizer blocks now accept initial design archive matrix to the designs input port. The matrix may contain either values of variables only (becomes a set of initial guess points), or both variables and pre-calculated objective and/or constraint values.


  • Fixed a bug in line plots which corrupted plots with added error bands in case if one of the error values is zero.

pSeven 2.1 Release Candidate 2

Release date: 14 February 2014.

This release includes MACROS 2.1 Release Candidate 2.


  • New pSeven installer with better Windows support.

pSeven 2.1 Release Candidate 1

Release date: 29 January 2014.

This release includes MACROS 2.1 Release Candidate 1.


  • Fixed Program in Windows version being unable to kill the process it created which could freeze the workflow if it contains an Program block with process timeout set.
  • Fixed non-existing links displayed in Edit after closing the link creation dialog without selecting the ports to link.

pSeven 2.0

Release date: 25 December 2013.

This release includes MACROS 2.0.


The final release of version 2.0 receives a new name, pSeven, which is now used throughout this manual.


  • pSeven for Windows now supports working with projects from network paths. To browse network paths, use the address bar in the Filesystem pane.
  • Redesigned the sidebar panes in Workspace for easier file management in projects. The sidebar now always shows the Filesystem and Project panes and there is no need to switch panes to add files to the project.
  • Added a multi-save dialog for the case when you close a project with unsaved changes in several workflows or reports. In this dialog you can discard all changes or select which ones you want to save.
  • Redesigned journal data analysis interface to speed up plotting and extracting data from workflow journals:
    • New Project database pane provides a direct view into project database down to the level of individual values.
    • Data series may now be created by simply dragging journal records to the new Data series pane. Multiple records may be converted to data series using Ctrl+Click selection.
    • Dragging journal records to main window area automatically creates a data table and adds all its columns to data series list.
    • In addition to manual configuration, plots may now be created by selecting one or more data series on the Data series pane and clicking a plot button on the toolbar.
  • Improved filtering in the link configuration dialog: the blocks to display can now be selected from a list, and window panels can be swapped to change source and target blocks.
  • Improved the configuration interface of TextGenerator and TextParser.
  • For more convenience, CSVParser now ignores trailing spaces by default, enabling it to parse files containing empty fields at the row ends correctly. The option to merge delimiters is now on by default; to switch back to the old behavior, set merge delimiters off.
  • Various minor usability improvements in Analyze.


  • Fixed Optimizer delaying the display of its logs in the console.
  • Fixed Program sometimes failing on process timeout (outputs False to the success port) even if the option to fail on timeout was canceled in block configuration.
  • Fixed the incorrect behavior of scatter and line plots in case when their source data contains NaN values.
  • Fixed a bug in pSeven Shell under Windows which made the commands containing Unicode symbols disappear from the command line history.
  • Fixed no error message appearing on an attempt to save a workflow or report into a non-existent directory.
  • Fixed the tiling layout of plots in Analyze taking up more area than is available on screen by default.

2.0 Release Candidate 2

Release date: 11 December 2013.

This release includes MACROS 2.0 Release Candidate 2.


  • Fixed a bug in Analyze which resulted in corrupted tables and plots after renaming data series on the Data series pane.
  • Fixed the inability to edit or run a workflow containing an older version of PythonScript block (as in workflows saved from pSeven versions prior to 2.0 Release Candidate 1).

2.0 Release Candidate 1

Release date: 26 November 2013.

This release includes MACROS 2.0 Release Candidate 1.


  • 3D plotting added in Analyze. pSeven now supports colorful 3D scatter plots supporting many advanced features:
    • Interactive plots with quick zoom and rotate functions.
    • Gradient point color mapping: preset multicolor gradients and single-color gradient with point color mapped to its index, useful for plotting 3D optimization trajectories.
    • Normal, log and square root axis scales which may be set for each axis independently.
    • Autorefresh support to update the plot in real-time while running a workflow.
  • Redesigned journal data analysis interface to speed up plotting and extracting data from workflow journals: - New Project database pane provides a direct view into project database down to the level of individual values. - Data series may now be created by simply dragging journal records to the new Data series pane. Multiple records may be converted to data series using Ctrl+Click selection. - Dragging journal records to main window area automatically creates a data table and adds all its columns to data series list. - In addition to manual configuration, plots may now be created by selecting one or more data series on the Data series pane and clicking a plot button on the toolbar.
  • Saved reports no longer depend on the main project database and are not corrupted even if project database is deleted (with the exception of plot and data series refresh functions which naturally stop working since there is no source to query data update from). Plot and table data (the data series) are now saved in report documents created in Analyze, not in the main project database. It also means that each report now works with its own data set.
  • Workflow editor improvements:
    • A workflow may now be copied to block exchange buffer and pasted into another workflow as a new block, making hierarchical design and workflow re-use much more intuitive and easier.
    • Better fit to size and workflow graph placement after zooming.
    • Cleaned up auto-layouting schemes.
  • New concept of working with variables in PythonScript blocks with added convenience and useful features:
    • Configured variables now have user defaults which serve as fall-back values in cases when the block is not able to read variable from port or the embedded script does not assign a value.
    • Default value becomes a block option and thus may be mapped to workflow run configuration. With this update, the feature of adding user-defined options in PythonScript blocks was removed as redundant.
    • Multiple variables with same settings and different names may be added at once.
  • Redesigned block sandboxes. Sandbox directory may now be created inside the project directory. Block configuration now defines the sandbox cleanup rules. See section Sandbox for more information.
  • File type variables in PythonScript are made compatible with Python file-like objects: they now mimic Python files behavior and methods.


  • Fixed incorrect handling of symbolic links in block sandbox directory which broke the sandbox cleanup. Now if sandbox cleanup rule is set, all sandbox contents are deleted properly, including symbolic links themselves; link targets are not deleted.
  • Fixed the opportunity to save a workflow outside the current project in Windows by saving it to another logical volume, which then leads to a workflow crash on execution.
  • Fixed opening console log files in Windows when there is no viewer associated with .log files; pSeven now shows an “Open with…” dialog.
  • Fixed incorrect pSeven Shell shortcut created in Windows Start menu.


Release date: 17 October 2013.

This release includes MACROS 1.11.1.


  • Fixed memory leaks caused by automatically refreshing plots.
  • Fixed automatically refreshing plots not refreshing at workflow finish, leaving the plot outdated until it is refreshed manually.
  • Fixed pSeven slowdown caused by console logs.
  • Fixed error message not appearing when trying to open a workflow which is not a part of the active project.
  • Fixed links sometimes displayed in the workflow despite link source or target port is already deleted.
  • Fixed reports not opening automatically when opening a project.
  • Fixed report not updating its name in pSeven if the report file was renamed manually.
  • Fixed the unexpected behavior of bar charts when source data contains NaN/Inf values. Such values are now ignored (not displayed).
  • Fixed the ability to overwrite an open workflow or report, and added a warning when trying to overwrite an existing file.
  • Fixed a crash in block configuration dialogs caused by rapidly clicking the Apply button.


Release date: 4 October 2013.

This release includes MACROS 1.11.0.


  • Added a button to reset zoom in Edit.
  • Added CATIA block integrating CATIA V5 models in pSeven workflows. The block is in development and is currently found in beta blocks in the Block Library.


  • Fixed workflow freezing sometimes when it received an exception from block, instead of terminating and reporting an error.
  • Fixed pSeven crashing when a license file is not present.
  • Fixed incorrect interpretation of empty journal names which resulted in data being appended to the previously specified journal after the journal name column is cleared. Now clearing a journal name in the Journaling pane will remove the watcher from the corresponding port.
  • Fixed the display of logarithmic scale plots in Analyze.
  • Fixed plot filling under line not working when the line is not solid (for example, dashed or dotted).
  • Fixed the value of GTOpt/GlobalPhaseIntensity option in Optimizer always being displayed as changed even if it is default.
  • Fixed CSVParser to correctly parse Unicode headers under Windows.


Release date: 19 September 2013.

This release includes MACROS 1.10.5.


  • Added error bars support to line and scatter plots in Analyze.
  • All blocks working with files will now preserve file open/close state. In particular, a block that received an open file will not close it before output, so the file keeps its seek position. Note that common blocks which create files (like File) output a closed file, so this feature is primarily intended for files output by PythonScript blocks.


  • Fixed a bug which lead to a workflow error when a few hundred files are opened in a loop.
  • Fixed File block crashing when an empty file is specified as the initial data source.
  • Fixed pSeven standard output not visible in the console when it is launched by an Program block from another pSeven instance.
  • Fixed some blocks working with files not closing the file before output when they should. All blocks now follow the logic described above (preserve the file state).
  • Fixed incorrect values displayed in tables including more than 100 rows.
  • Added NaN and Inf value filtering to plots. Data containing NaN/Inf values will no longer break the plot, though such values will not be displayed.
  • Corrected name validation when grouping blocks.
  • Corrected error message when trying to open a project in pSeven version older than the project version.


Release date: 3 September 2013.

This release includes MACROS 1.10.4.


  • pSeven now creates scheduled backups of the project database while there is a workflow running. If the project crashes, the database will be restored from a recent backup copy automatically.
  • Added new plot types in Analyze: scatter plots and bar charts.
  • Reworked the plot configuration dialog.
  • Added column sorting and manual column ordering in tables on the Analyze screen.
  • Added log message type highlighting in the console.


  • Fixed Optimizer validation error when variable or constraint bounds are mapped to port.
  • Fixed certain combinations of cell delimiter, decimal delimiter and escape character not working in CSVParser configuration.
  • Fixed the inability to create links inside a block group if one of the grouped blocks has the same name as the group itself.


Release date: 21 August 2013.

This release includes MACROS 1.10.3.


  • More efficient project database scheme allowed to significantly reduce the database size and memory consumption when journaling large data.


    Unfortunately, this update required some changes in the database structure which make it incompatible with previous versions. When upgrading existing projects to 1.8.3, the data stored in workflow journals will be removed, which also breaks the saved reports. Saved workflows and all other project settings, such as run configurations, are not affected and will be upgraded normally. See Compatibility Issues section below for details.

  • Project database files (project.p7pdb) no longer have the hidden attribute and are visible in the file browser at the Workspace screen (but can not be deleted using pSeven file operations).

  • Added the block copy/cut/paste functionality which allows copying a block to the clipboard and paste it in another workflow, preserving the block configuration.

  • Added placement guides in the workflow edit area on the Edit screen (shown when you move blocks).

  • Added the Open log button to the pSeven console to easily view and copy long logs without using mouse selection in the console.

Compatibility Issues

  • The 1.8.3 release changes the project database scheme (project.p7pdb file format). This change is a result of an update which makes the database more efficient, reducing its size and memory usage when journaling large data. Unfortunately, the new format is not compatible with old database files, and due to this the database contents will be lost when upgrading older projects to 1.8.3.

    In practice this results in emptying the saved journals when a project from version 1.8.2 or older is upgraded to 1.8.3 or newer versions. Other project settings remain intact. Workflows in the upgraded project are fully functional, workflow design and configurations are not affected. Workflows may be re-run after the upgrade to restore the journal data, since journaling settings are saved in the workflow run configuration.

    Before upgrading to 1.8.3, it is recommended to backup your projects and to export the report data (tables) to CSV files.


  • Fixed Qt library conflicts which lead to pSeven crashing when it is used on a Linux system with its own set of Qt libraries (for example, with KDE installed).
  • Fixed plot refresh function in Analyze not working if the plot uses data from a journal with non-default name (specified by user).
  • Fixed wrong Fit to size button behavior in Edit (if the workflow contained only one block, always zoomed out when clicked).
  • Fixed the console log not scrolling automatically when executing a Python script in pSeven.
  • Fixed Optimizer crashing on start if different hints were set for the components of a vector constraint.
  • Fixed incorrect value conversion in TextParser: if the read operation converted a value to scalar and it was output to a vector port, the resulting output had scalar type, not vector. Now on-read conversion will not override the port type.
  • Fixed TextParser locking an input file of project origin in Windows, so the file could not be renamed or moved until you close pSeven.
  • Fixed the code on the Script tab in TextGenerator and TextParser not updating when changes are done on the Operations tab.
  • Fixed File showing wrong information on the Data initialization pane if the data was loaded and then initialization type was switched to “None”.


Release date: 7 August 2013.

This release includes MACROS 1.10.2.


  • pSeven for Windows now supports the 4-gigabyte tuning (4GT) feature. On 64-bit editions of Windows, it means that pSeven has 4 GB of virtual address space available, leading to better resources utilization on systems with large amounts of RAM. On 32-bit Windows editions, this feature also increases the address space up to 3 GB, if the feature is enabled in Windows boot configuration.


  • If some block options are added to the workflow configuration, and the block is removed from the workflow afterwards, the options are now also removed from the configuration as it should be.
  • Fixed Saver blocks always writing to the first file received, when several files are sent to the block on different steps.
  • Fixed no warning appearing when trying to create a project with a name that already exists.
  • Fixed the possibility of creating a workflow without configuration in pSeven Shell, which then blocks the workflow selector if opened in pSeven.
  • Fixed the incorrect behavior (unexpected disable/enable) of the Save report button in Analyze.
  • Fixed workflow validation error if a link is selected before validate.
  • Fixed Analyze not closing after the project is closed.


Release date: 23 July 2013.

This release includes MACROS 1.10.1.


  • Added Kompas3D block to integrate KOMPAS-3D models in a pSeven workflow. This block is at an early development stage, providing only the most essential functionality, and is currently found in beta blocks in the Block Library.

  • Added the support for Parasolid (.x_t) and ACIS (.sat) export formats to SolidWorks blocks.

  • Added basic support for integer variables to Optimizer blocks (select the integer type when adding a variable).


    This functionality in MACROS is still experimental, so the same limitations apply to Optimizer in pSeven. In particular, integer variables are not supported in multi-objective and/or constrained problems.

  • Projects are now opened by double-clicking them in the file browser pane.

  • Removed the Create new workflow and Create new report buttons from the Workspace screen. To create workflows and reports, use the File menu or New buttons on the Edit and Analyze screens, respectively. Both screens are now available right after creating or opening a project.

  • Data viewers in Analyze now allow changing their window titles.

  • Double-clicking a data series in Analyze now automatically opens a data viewer.

  • The console now opens automatically after entering any command.

  • Added tooltips for beta blocks in the Block library pane.


  • Fixed existing watchers being removed when adding a block option to the workflow configuration.
  • Fixed the inability to copy and move files between different disk partitions using the built-in file browser.
  • Fixed renaming objectives in Optimizer blocks leading to crash at the workflow start if sparse gradients are used in the problem.
  • Fixed errors in handling the smoothing option in ApproxPlayer blocks.
  • Fixed CSVParser blocks producing an error when selecting empty lines in the file contents preview.
  • Fixed TextGenerator and TextParser blocks clearing the operations list on clicking Apply.
  • Fixed TextGenerator and TextParser blocks always showing Undefined type for output variables.
  • Fixed the ability to set a negative value for the cluster/advanced/retry_count option in Program blocks which could crash pSeven.
  • Fixed File blocks always requiring an input to the do port in case when the block has to be started more than once in the same workflow run (for example, when it is working inside a cycle).
  • Fixed Random blocks not saving their configuration.
  • Added a check if a report exists before creating data tables in Analyze.
  • Fixed the inability to uplink several ports with compatible types to the same port automatically.
  • Fixed pSeven Shell creating an invalid configuration file, which lead to an inability to start pSeven until the configuration is deleted in case when pSeven Shell was run right after installation.
  • Fixed some memory leaks.
  • Fixed pSeven not starting in some old Linux distributions satisfying the new minimum system requirements (see section System Requirements and version 1.7.6 changelog for details on the updates that uncovered this issue). This is a GUI-specific bug, core functionality and pSeven Shell were never affected.


Release date: 14 July 2013.

This release includes MACROS 1.10.0.


  • Better support for automatic uplinking of multiple nested block ports to the same Composite block port.
  • Workflows in a project are now always sorted by name, not in the creation or opening order.


  • Fixed TextParser and TextGenerator blocks not displaying the template preview when the template is loaded as a project file.
  • Fixed Calculator converting NumPy scalar values to matrix.
  • pSeven now offers to save changes to reports on exit.
  • Plot order in a report is now saved.
  • Fixed several bugs in pSeven which caused memory leaks when opening block configuration dialogs.
  • Fixed incorrect display of the Dict data type in data editors.
  • Fixed inability to display NaN values in data editors.
  • Fixed copy and paste text not working properly in the new project dialog.



Release date: 26 June 2013.

This release includes MACROS 1.9.6.


  • Reviewed the block sandbox logic, resulting in a more clear and well-defined sandbox behavior.


  • Fixed pSeven installer quoting the value of the DATADVD_LICENSE_FILE environment variable which could lead to the license manager being not able to find the license under Windows 7.
  • Fixed pSeven Shell under Windows crashing when a running workflow is terminated by keyboard interrupt.
  • Fixed integer overflow problems in data editor: integer values are now validated upon input.
  • Fixed problems with encoding in exception messages containing Unicode symbols (for example, file names in native language).
  • Fixed name clashes when ungrouping blocks.
  • Fixed an incorrect handling of the case when user specified a wrong script interpreter in Program.
  • Fixed Program not removing its temporary files in Windows.
  • Fixed zero values displaying in TextParser as empty fields.
  • Fixed the sorting of data series on the Analyze screen.
  • Fixed exception when trying to save an empty plot as an image.
  • Various minor fixes in pSeven interface dialogs.


Release date: 14 June 2013.

This release includes MACROS 1.9.5.


  • Since this release, pSeven uses a new build system. This is a very important upgrade since the new build system is specifically set up to support older Linux distributions. It removes all non-essential dependencies and thus increases the number of compatible Linux platforms — see section System Requirements for details.
  • Added the LSF cluster support to Program blocks.


  • Fixed workflow run configuration not applying option values selected from a drop-down list.
  • Fixed CSVGenerator not showing the output file name when reopening the block configuration dialog.
  • Minor bugfixes in code editors (PythonScript, Program and other scripting-enabled blocks).
  • Fixed the Esc key not working correctly in pSeven dialogs. Now works as intended: in secondary dialogs, pressing Esc is the same as clicking the Cancel button; in main block configuration dialogs, Esc does no action.


Release date: 28 May 2013.

This release includes MACROS 1.9.4.


  • Implemented cluster queuing support in Program. The block can now submit jobs to a cluster managed by SLURM or TORQUE.
  • Added the refresh function to the Data series pane on the Analyze screen.
  • Console Clear button will now remove all messages from the log instead of the old terminal-like behavior (adding empty lines and scrolling the log).


  • pSeven will no longer crash when terminating a workflow containing an SolidWorks block.
  • Fixed pSeven crashing when console input contains a symbol in the encoding different from the system default.
  • pSeven will now block user input to a dialog window when it performs time-consuming operations in background, to avoid crashing as a result of a fast action sequence which can not be handled in GUI.
  • Fixed PythonScript configuration dialog suddenly closing when the Enter key is pressed to insert a new line in the script.
  • Fixed several bugs in Program block type implementation which could lead to instability.
  • Fixed the Autolink dialog sometimes appearing on a double-click on a link (intended double-click behavior is to open the Links connectivity dialog).
  • Links connectivity dialog fixes:
    • Fixed the dialog being slow when displaying a few dozens of ports.
    • Fixed incorrect port compatibility check which allowed user to connect incompatible ports.
    • The dialog will now show correct block names when creating an uplink.
  • Project upgrade fixes:
    • When opening an older project which requires an upgrade, the project will now actually open after the upgrade is done (before the fix, the upgraded project had to be opened manually).
    • Upgrade will no longer fail if a project database file is not found in the project to be upgraded.
  • Fixed block sandbox path change being ignored when checking for changes in workflow (for example, when user closes an unsaved workflow).


Release date: 13 May 2013.

This release includes MACROS 1.9.3.


  • Added the capability to explore the workflow hierarchy on the Run screen.
  • Added refresh and auto-refresh functions to the plot and data viewers on the Analyze screen.
  • Added detailed tooltips for all non-beta blocks in the Block Library (hover the block typename on the Block library pane to show).
  • Added tooltips displaying a full list of compatible types for ports on the Ports tab of the block properties window.
  • Added the support for date, time, hostname and process ID placeholders in Program substitutions.
  • The stdout and stderr messages from Program blocks are now also displayed in the console (info and warning loglevel, respectively), in addition to being sent to the block’s output ports.
  • pSeven will now inform the user when performing long operations during which the interface becomes unresponsive.


  • Fixed links disappearing when one of the linked ports is renamed.
  • Added journal locks which are active during a workflow run, to prevent accidental removal of a record in the journal which previously lead to workflow crash.
  • Fixed several bugs which caused the Journaling pane on the Run screen to lag when there are many (about 100 or more) journaled ports in the workflow.
  • Fixed TextGenerator not being able to apply C format to vector variables.
  • Fixed incorrect display of strings containing XML tags in value editors.
  • Fixed temporary files not being removed from a block sandbox.
  • Prohibited adding special symbols to the names of data series and journal records when renaming them, which previously lead to pSeven not being able to manipulate the contained data anymore.
  • Corrected the error message displayed when the user attempts to save a workflow outside the project’s directory.
  • Corrected project upgrade log messages which sometimes were ambiguous.


Release date: 6 May 2013.

This release includes MACROS 1.9.2.


  • Added block autolinking feature: if a link is added by dragging link handle from block to block, and on both sides there are ports with same names and compatible types, pSeven will now offer a simplified linking dialog with linkable ports connected by default.


  • Fixed precision loss when exporting data from the dataset viewer to a CSV file.
  • Fixed unexpected removal of all links connected to an input port when the user removes an uplink connected to this port.
  • Fixed broken selector for the number of rows in a CSV file preview displayed by CSVParser blocks.
  • Fixed Submatrix block always demanding an input to its do port.
  • Fixed several misbehaving interface elements.


Release date: 15 April 2013.

This release includes MACROS 1.9.1.


  • Added a mechanism to upgrade projects from previous pSeven versions to the current version. Project upgrade invokes automatically when opening a project from an older pSeven version.
  • The SolidWorks block has been released.
  • Block type will be now shown on the block icon in the workflow.
  • pSeven Help will now open in its own window, allowing to switch between reading the manual and editing a workflow without closing Help.
  • Added an option to data viewers to select the line ending type for the exported CSV files.
  • pSeven will now show a license expiration warning two weeks prior to the license end date.


  • Fixed a crash when saving or loading files with very long file- or pathnames. Such paths can not be handled by pSeven internal database, and a proper warning will be issued in this case. Also added a note to FAQ.
  • Fixed futile attempts to set a terminal codepage when pSeven is started in GUI mode under Windows.
  • Fixed a bug in logging system timers setup which could lead to non-breaking errors being produced into the console log after starting a workflow.
  • Fixed block names in the workflow disappearing sometimes.
  • Various visual fixes and small improvements in pSeven interface.


Release date: 2 April 2013.

This release includes MACROS 1.9.0.


  • Data series may now be exported to a CSV file.
  • Improved console logging. The console will now save a full log since session start.
  • Better support for low screen resolutions.


  • Fixed wrong behavior in case when the user removes a data series source (a journal or a journal record) but does not remove the data series itself.
  • Corrected pSeven behavior when attempting to rename/remove an open project or open a non-existing project.
  • Fixed the file lock not being removed from a project after closing it in Windows.
  • Fixed an TextGenerator block not creating the output file when generation script is empty. The block will now copy template to the output as expected.
  • Fixed a bug which caused a crash when switching to the Run screen after a port with a watcher placed on it was removed from a block, or a block with watched ports was removed from the workflow on the Edit screen.
  • Fixed the data from the root output ports not being recorded to workflow journals.
  • Fixed selected data series not showing up on the Advanced tab of the plot properties window.
  • Fixed the inability to set an empty string as a value of StringScalar.
  • Corrected several visual artifacts and inconveniences in the graphical user interface.


Release date: 13 March 2013.

This release includes MACROS 1.8.5.


  • The graphical user interface for pSeven was completely redesigned. In addition to the new visual appearance, it also adds several new features and many fixes both in the general interface and in block configuration dialogs:
    • Added project view mode in the built-in file browser on the Workspace screen.
    • Added workflow outline (tree view showing the block hierarchy) on the Edit screen.
    • Added block options configuration pane on the Run screen (options may now be specified in the run configuration, see below).
    • Significant interface changes on the Analyze screen due to the new concept of data analysis with pSeven.
    • Enhanced the link creation dialog. If a link is created by dragging it between blocks, only these blocks will be shown in the dialog. The user may also manually filter blocks and ports shown in the dialog by using built-in filters.
    • Made option dialogs more convenient: non-default option values are highlighted, added option reset functions (the user no longer needs to reset options to default by manually entering the default option value).
    • Changed the rules of automatic block layouting in workflow editor.
    • Completely redesigned the configuration dialogs for Const, Submatrix and Switch blocks.
    • All other blocks of the release type also have their configuration dialogs changed to improve usability.
  • Increased graphical user interface performance at workflow run-time.
  • The graphical interface is now the default mode when pSeven is started from the command line.
  • Introduced the projects concept to pSeven. A project is a portable collection of workflows and reports sharing one common data storage. Consequently, a workflow is no longer a separate entity — it always exists within a project.
  • Projects may be opened from a Windows network path when current working directory is set to this path.
  • Workflow results analysis is now done within the framework of pSeven reports. Reports are a new kind of document replacing the previously used datasets.
  • Reworked the database query mechanism for reports. In particular, this update enables processing significantly greater amounts of data generated by the workflow.
  • Starting with this release, pSeven projects, saved workflows and reports will be compatible with newer pSeven versions. Compatibility on the scripting level will be added in one of the nearest releases.
  • Revised the workflow start and run logic which made workflows to start much faster and consume less RAM at start.
  • All blocks have been divided into two large groups: release (stable) blocks and beta blocks. See Block Library for details.
  • Regrouped blocks in the Block Library to reflect the above change.
  • Each block type is now a part of one of the block packages coming with pSeven. Default install currently includes three such packages: std (standard blocks), macros (the blocks integrating MACROS Generic Tools into pSeven) and adapters (the package for blocks allowing pSeven integration with third-party applications).
  • Block naming rules were changed.
  • All blocks capable of working with files now have the same file handling capabilities as the File block. In particular, this allows for resource files uploaded to blocks to be saved with the workflow.
  • Added block sandbox which serves as a safe working area for a running block. In particular, it allows transferring files between blocks at the run-time.
  • Added environment handling for individual blocks.
  • Potentially unstable feature of mapping block options to ports has been replaced with a simpler and more effective approach to change block options for a single run. Option values for all blocks in the workflow may now be specified in the workflow run configuration.
  • Added do and done signal ports to all blocks.
  • CSVParser and CSVGenerator blocks now support different input and output file encodings.
  • Minimized the dependencies of the integrated Python interpreter.
  • Integrated NumPy was updated to version 1.7.0.
  • Added the support for NaN values in pSeven data viewers.


  • Changes in block configuration are no longer applied on the fly, and configuring a block no longer changes the workflow until you accept the block configuration. This also allows you to properly cancel unwanted configuration changes.
  • Links will no longer reset their positions to default when blocks are moved on the Edit screen.
  • Fixed many minor GUI errors in block configuration dialogs.
  • Fixed errors on script import/export for Program blocks.
  • TextGenerator and TextParser blocks will no longer replace removed operations with a pass line in the script view.
  • Fixed nested blocks (blocks contained inside an Composite block) not appearing in the run statistics window when the workflow finishes.
  • Fixed inability to open a directory with the name containing whitespaces at the end in the built-in file browser.
  • Fixed an empty pane sometimes displaying instead of the file tree in the builtin file browser.
  • Fixed non-existing link showing in the workflow when the user decided to cancel the link creation after dragging it to the target block.
  • Fixed pSeven crash when deleting several linked blocks from the workflow simultaneously.
  • Fixed wrong default name offered when saving a workflow under Windows.


Release date: 21 January 2013.

This release includes MACROS 1.8.3.


  • SolidWorks block improvements:
    • Added support for the SolidWorks Simulation add-in.
    • Added support for assemblies, configurations and sensors.
    • Added capability to select SolidWorks version to use if several different versions are installed.
  • TextGenerator block now supports Fortran and C-style string formatting.


  • Under Windows, different versions of pSeven installed simultaneously will now be visible as separate versions in Add/Remove Programs.


Release date: 19 December 2012.

This release includes MACROS 1.8.1.


  • Improved DataAccess.FileLoader and DataAccess.FileSaver blocks to support the same functionality as implemented by the DataAccess.File block.
  • Added an option to merge delimiters (ignore repeated delimiter symbols) to CSVParser block.
  • Block Library improvements:
    • Pure hierarchical structure replaced by flexible block grouping system with tag support.
    • Each block type now has a tooltip with brief description and a Help button linking to block’s documentation in the manual (hover the blocktype name with your mouse to bring up the tip).
  • Various GUI usability improvements.


  • Fixed block validation displaying an empty validation window when no block configuration errors are found.
  • DataGenerators.Const blocks no longer allow user to create constants of Blob type.
  • Fixed DataAccess.CSVParser not working when the columns to import are specified as a “:-1” slice.
  • Fixed duplicated data appearing at the info port of an DRCompressor block.
  • Fixed plot windows ordering (tile, cascade) on the Analyze screen.
  • Fixed disappearing block toolbar on the Edit screen.
  • Fixed the plot refresh button becoming inactive if another plot designer window was opened and closed.
  • Typesystem fixes:
    • Scalar values and File now allow conversion to a single-element List as they should.
    • Properly prohibited all Dict conversions.
    • Improved conversion from Python iterables to StringVector and StringMatrix.
  • Multiple minor GUI fixes in block configuration windows.


Release date: 3 December 2012.

This release includes MACROS 1.7.7.


  • New block SolidWorks to integrate SolidWorks with pSeven.
  • Python scripts now can be run right from the pSeven graphical interface.
  • Implemented the configuration dialog for File block type.


  • Fixed the floating license check so the user will be able to run workflows even if the connection to the license server is lost after starting pSeven.
  • Fixed various issues appearing in the run configuration if it was saved from pSeven Shell and opened in pSeven.
  • Fixed option validation for the Program.TextGenerator and Program.TextParser blocks.
  • Fixed the script editor in the Program.TextGenerator and Program.TextParser blocks removing parentheses from the script.
  • Fixed the “suspicious unknown environment variable” message appearing in a terminal on pSeven start.
  • Fixed a Windows-specific bug which could freeze pSeven after opening a block configuration dialog.
  • Fixed GTIVE technique selection dialog to display only the techniques available in the selected block mode (sample- or blackbox-based).
  • Fixed automatic case naming for the FlowControl.If block.


Release date: 22 November 2012.

This is a bugfix release which includes MACROS 1.7.6.


  • Fixed a bug which disabled the file select button in the DataAccess.CSVParser block configuration dialog.
  • Fixed sending wrong run configuration identifier to a workflow which resulted in pSeven crash when the workflow is run.


Release date: 20 November 2012.

This release includes MACROS 1.7.6.


  • DoE now supports adaptive DoE.
  • More useful exception reports from PythonScript blocks: now include true traceback for user script.
  • Watchers may now be added to all input/output ports at once.
  • If no journal name was set for a watcher, the watcher will now write data to a default journal.
  • Improved performance of CSVParser blocks.
  • If user specifies an input value of some Python built-in type in the run configuration in pSeven Shell, the value is now instantly converted to a pSeven data type following the rules described in section Python to pSeven Types Conversion.


  • Watchers configured in pSeven Shell will now save relative journal path and journals will be written correctly when the saved workflow is moved to another directory or host.
  • Watchers placed on the output ports of Composite blocks will now convert data types correctly before writing the data to journal.
  • Fixed save and load functions not working when there are Unicode symbols in path.
  • Fixed creating new workflow after an attempt to open corrupted workflow file.
  • pSeven Shell command will now return a configuration object as expected.
  • Fixed bugs in run configuration that caused inability to add watchers to ports.
  • Fixed error when watchers are set on both container (for example, root) block port and a block port uplinked to it.
  • Fixed uplinking ports when some of the uplinks already exist.
  • Fixed FlowControl.Switch block configuration dialog.
  • Fixed Scripting.PyScript block starting even if user-defined ports are not connected.
  • Fixed the Recent workflows menu not saving the recent workflows in some cases.
  • Fixed adding scalar outputs to DataGenerators.Const blocks.
  • Fixed warnings about an already opened file generated by the file preview in DataAccess.CSVParser blocks.
  • Fixed the display of constraint type hints in Optimizer block configuration dialog.
  • Fixed pSeven uninstaller removing paths not related to pSeven from system **PATH**.


Release date: 07 November 2012.

This release includes MACROS 1.7.5.


  • Added capability to import a CSV file to dataset.
  • CSVGenerator blocks will now start when no CSV header is specified.
  • pSeven Shell interface and the console in pSeven now feature a object which is an alias for the most recently used workflow (in pSeven this is workflow currently opened on the Edit screen).
  • Added status port to ApproxBuilder, DFBuilder, DRBuilder, IVE, and Optimizer blocks. The port outputs the block stop status and allows handling block exceptions in user interface.
  • Improved file write functionality of the Program.TextGenerator block type.


  • Fixed journal browser showing incorrect dimensions of matrices contained inside a list.
  • Fixed model info display bug in ApproxPlayer.
  • DataGenerators.RandomGenerator type blocks will work correctly when set to negative binomial or geometric distribution with probability parameter equal zero.
  • Fixed a memory leak.
  • Fixed a bug in Scripting.PyScript block GUI that sometimes blocked input to the script editing area.
  • Made grouping function unavailable if the selection includes a bound of the root block to stop it trying to “group” the bounds.
  • Fixed various minor bugs with zooming and selections.
  • Fixed the pSeven process freeze when pSeven is closed while a workflow is running. Workflow states are now checked on exit.


Release date: 23 October 2012.

This release includes MACROS 1.7.4.


  • Implemented block validation routines for all blocks.
  • Added automatic port uplinking support (single-click creation of new ports for a Composite block and linking them to selected input and output ports of its nested blocks).
  • Added support for placeholders in journal names specifying the workflow name, PID, hostname, date and time.
  • Added Robust Optimization support in Optimizer.
  • Added the capability to export models loaded into ApproxPlayer, DRCompressor, and DRDecompressor blocks at design-time to the Octave format.
  • Added script import/export functionality in the configuration windows of the TextGenerator and TextParser blocks.
  • Added an option to import Python modules into a FlowControl.If block.
  • Added sender selection to the console. Only the messages from selected senders (blocks) will be logged to the console.


  • Fixed option editors for options that have a predefined set of values not showing the list of values when configuring a block.
  • Fixed a bug which lead to workflow remembering a non-existing path as its save path, if user attempted to save new workflow (which was never saved before) to a non-existing path.
  • Fixed adding watchers to the workflow run configuration if the workflow was not saved yet.
  • Fixed input language change in Windows hanging pSeven.
  • Fixed a bug with workflow configuration not updating when user specifies new input value for a root input port.
  • Fixed first run of pSeven consuming more than 1 GB of RAM.
  • Fixed wrong conversion of the string representation of Inf/NaN values when they pass through a port.
  • Fixed saving multiple opened workflows on exit (the Save all button).
  • Fixed defining variables and variable grouping in the Optimizer block configuration window.
  • Fixed the Add option dialog window of the PythonScript block.
  • Fixed the Expand button for a Composite block.
  • Fixed reloading the template file in a Program.TextGenerator block.
  • Fixed the Program.TextParser block not running if there are unused variables in the script.
  • Fixed the IPython help() command not working properly in pSeven Shell.
  • Fixed installation of pSeven to a path containing whitespaces.
  • Fixed the dataset viewer not updating the dataset if it was edited while the viewer is open.
  • Fixed workflow run/stop button disappearing when user switches workflows while on the Run screen.
  • Fixed moving/copying files and directories to an existing path (file/directory overwrite) in the pSeven file browser.
  • Fixed the table sorting in GUI resetting after table data update.


Release date: 9 October 2012.

This release includes MACROS 1.7.3.


  • Automatic block layouting in workflow editor.
  • Improved dataset editor interface.
  • Created plots are now saved with the workflow.
  • Plots and datasets are now automatically updated when journal or dataset data changes.
  • Improved data editors and viewers.
  • Improved run configuration interface.
  • When stopping the workflow, user now has an option to wait until currently active blocks finish execution or to terminate the workflow at once.
  • Improved error reporting for the Program.TextParser and Program.TextGenerator blocks.
  • Improved Program.ProgramRunner block performance and stability.
  • Installer now offers to add pSeven to system PATH so it will be available for all users after installation.
  • Switched to the FlexNet licensing engine (see the License Setup chapter).


  • Fixed installation error when installer is run on a system with no PATH environment variable.
  • Fixed the installation of pSeven over an older version.
  • Fixed a memory leak.
  • Multiple fixes in the configuration GUI for the ApproxBuilder, ApproxPlayer, Optimizer, TextParser and TextGenerator blocks.


Release date: 19 September 2012.

First public release. Includes MACROS 1.7.2.