Datanami
EnterpriseAI
HPCwire Japan
QCwire
HPC & AI Wall StreetJan. 20, 2020 — Appentra is releasing Parallelware Trainer 1.4, stating it greatly improves the deployment and user experience in supercomputers by allowing configuration through environment variables.
From the release notes:
New in Parallelware Trainer 1.4
Environment variables can now be used to specify a custom license path as well to set default analysis flags.
This has been a popular request, especially for supercomputer deployments where a Modulefile can take advantage of these variables to specify the path to a shared license file as well as to add include paths to non-standard locations. One common example of the latter is adding the path to the MPI header files. We have found that some MPI Modulefiles modify gcc include path environment variables to add the headers location. As a result, besides our new environment variables, Parallelware Trainer also now takes those gcc environment variables into account.
Detailed information about these variables is available in the Server installation section of the user manual and the Customizing Parallelware Trainer through environment variables blog post.
New Reanalyze action to force analysis.
The new Reanalyze action allows the user to manually force a new analysis of the file to search for opportunities, defects and recommendations. It is accessible from both the File menu and the file tab context menu.
File and version management improvements.
Files and folders can now be created, renamed, copied and deleted from the Project Explorer context menu.
Version management has been enhanced by creating new actions in the File menu to create new versions as well as restoring and exporting individual versions in a simpler way. Furthermore, the version source code can no longer be modified through the version manager.
Other minor changes.
Bug fixes and other minor changes such as file tab name disambiguation. Take a look at the full changelog through Help > About > What’s new?.
Try Parallelware Trainer for free here and make code parallel today.
Source: Appentra
