Releases: extra-p/extrap
Releases · extra-p/extrap
Release list
Extra-P 4.2.5 Release
Extra-P 4.3.0 Beta 2 Release
Enhanced GPU measurement import
- Uses Score-P call-side IDs to map kernels into the CPU call tree
- Import plugin (experimental) for Nvidia Nsight System databases and Nsight Compute reports
- Extra-Profiler for simultaneous CPU-GPU profiling (currently Nvidia GPUs only)
- Import plugin for Extra-Profiler profiles
Comparison of experiments
- Experiment comparison assistant
- Automatic call-tree mapping
- Difference models
- Visualization of differences as a heat map
- Dynamic comparison annotations
- Projection of expected performance based on the roofline model
Enhancements
- Improved formula rendering
Full Changelog: v4.2.4...v4.3.0-beta2
Extra-P 4.3.0 Beta 1 Release
Enhanced GPU measurement import
- Uses Score-P call-side IDs to map kernels into the CPU call tree
- Import plugin (experimental) for Nvidia Nsight System databases and Nsight Compute reports
- Extra-Profiler for simultaneous CPU-GPU profiling (currently Nvidia GPUs only)
- Import plugin for Extra-Profiler profiles
Comparison of experiments
- Experiment comparison assistant
- Automatic call-tree mapping
- Difference models
- Visualization of differences as a heat map
- Dynamic comparison annotations
- Projection of expected performance based on the roofline model
Enhancements
- Improved formula rendering
Full Changelog: v4.2.4...v4.3.0-beta1
Extra-P 4.2.4 Release
Bug Fix
- Fixes bug when using Python 3.13, where Extra-P crashed during startup
- Fixes bug in deserialization of experiments with segmented models
Extra-P 4.2.3 Release
Bug Fix
Fixes bug when importing cube files with scaling type weak threaded.
Extra-P 4.2.2 Release
Bug Fixes
- Updates the pyCubexR dependency to include the newest bug fixes.
- Fixes a bug that caused storing the wrong raw values in the experiment when keep values is selected
- Fixes a bug in the TreeView that in some cases caused a crash of the GUI.
Extra-P 4.2.1 Release
Bug Fix
Fixes bug in modeler widget that prevents modeling with measures other than the mean
Extra-P 4.2.0 Release
Scaling analysis support
- Automatic detection and conversion of strong scaling
- Improved strong scaling support for CUBE files
- Enhanced weak scaling modeling for CUBE profiles of multi-threaded applications
More options for selecting the measure of aggregation
- Modeling the minimum and maximum of the measured point is now possible
New modelers
- Segmented modeling: modeling of two behaviors in one model
- Support for segmented performance models with a single or multiple model parameters.
- Extension of the available plotting features to support segmented models.
- Adaptive modeling: optionally uses a neural network to improve the noise resiliency of the models
Measurement point suggestions based on Gaussian Process Regression
- The GUI features a view for measurement point suggestions.
- Several suggestion approaches help the user to complete or extend the set of available measurement points
to improve model accuracy further and enhance the noise resilience of the model while optimizing modeling cost.
Enhancements
- Functions can be formatted in Python and Latex format
- Function names are shortened in plots
- Keeps track of number of repetitions
- Optionally keeps all measured values in the experiment
Full Changelog: v4.1.2...v4.2.0
Extra-P 4.2.0-beta.1 Release
Scaling analysis support
- Automatic detection and conversion of strong scaling
- Improved strong scaling support for CUBE files
- Enhanced weak scaling modeling for CUBE profiles of multi-threaded applications
More options for selecting the measure of aggregation
- Modeling the minimum and maximum of the measured point is now possible
New modelers
- Adaptive modeling
- Segmented modeling
Measurement point suggestion
- Measurement point suggestions based on Gaussian Process Regression
Full Changelog: v4.1.2...v4.2.0-beta.1
Extra-P DEEP-SEA Release 1
Comparison mode
- Comparison assistent including call-tree mapping
- Projection of expectations
- Interactive comparison annotations for comparing experiments
- New plots to assist in comparison
- Updated existing plots for basic support of comparison
Loading of joint CPU-GPU profiles
- From Extra-Profiler files, including overlap
- Support for CUBE profiles of GPU applications (including call-side information)
New analyses and aggregation
- Agregation of models using sum and max operator
- Analyses for parallel efficiency and
Full Changelog: v4.1.2...DEEP-SEA-V-1