extrap
extrap copied to clipboard
extra-p
Extra-P, automated performance modeling for HPC applications
Traceback (most recent call last): File "C:\Users\Marcus\AppData\Local\Programs\Python\Python38\Scripts\extrap-script.py", line 11, in load_entry_point('extrap', 'console_scripts', 'extrap')() File "c:\users\marcus\git\extrap\extrap\extrap\extrapcmd.py", line 194, in main text = format_output(experiment, printtype) File "c:\users\marcus\git\extrap\extrap\fileio\io_helper.py", line 133, in format_output text...
Under the readme publications, there are 3 listed. But there is at least a fourth one, the foundational paper! The youtube video mentioned in the readme has even more publications...
I will disable the test for the adaptive modeler in the new release for now. This has the following reasons: - Test takes 7 min. to pass, see the output...
I would remove all tests that test like the following: - Regex against string output of the modeler, where the model is created either by reading data from input files...
Could design and add same features as in GUI to the command line tool.
Dear extrap team, with extrap-gui from extrap 4.1.2 I tried to load from the command line as either of the following way: ``` extrap-gui --cube case_working+profile --scaling strong extrap-gui --scaling...
The segmented modeler could benefit from the following extensions: - Support for multiparameter models (need to add segmented model support in the multi parameter modeler) - Support for several change...
When saving the results of a model created with the segmented modeler, when reloading using open experiment the following error comes up: Commit 1945ffe of Branch #[v4.2.0](https://github.com/extra-p/extrap/tree/v4.2.0) `$ Traceback (most...
Downloading the 4.2.0 tarball and calling `pip install .` leads to ``` ERROR: Could not find a version that satisfies the requirement PySide6-Essentials~=6.4 (from extrap==4.2.0) (from versions: none) ERROR: No...
I suggest adding these two functions to the `unique_list.py` file. It will make the UniqueList picklable and thus enable parallelization, e.g., with multiprocessing or concurrent.futures. ``` def __reduce__(self): return (self.__class__,...