zadorlab
Pynta restart for NERSC and Polaris
Previous comment from Ray (October 2023): I have updated the way we map tasks on each node for ALCF machines. Each task runs on a different FWorker, and each FWorker is associated with a node. This is available for multilauncher. The optimal approach is to set num_jobs to the number of nodes.
Additionally, I have added functionality to use PWDFT, including the calculator and related functions.
Updated comments from Shinae (Feb 2024)
I have updated the way we restart Pynta based on how Ray implemented the restart from Polaris.
From Pynta object, machine=<Machine type> should be specified to restart Pynta. Also for NERSC and any other machines, workflow id should be added to pyn.reset() to rerun the previous workflow.
pyn.reset() is tested in Perlmutter and was able to restart from queue=True mode.
This updates include Trevor's pull request (https://github.com/zadorlab/pynta/pull/33), which not yet merged.
Updated comments from Shinae (March 2024) I rebased to current master, with recent changes. Added @rayhe88 as the author. I still need to add him as an author for previous commits.
