Closed ye-luo closed 3 days ago
Test this please
It seems that I made a fix to VMC but there is actually no change to VMC. DMC is the one got affected. Due to branching, coordinates in ParticletSets may not be the same as Walkers. Using the former one for restart is incorrect.
I was wondering about any similar Walker / ParticleSet manipulations that might be lurking in the DMC sources. There are reports of the batched code paths being noisier than legacy, so perhaps there is some similar confusion.
I was wondering about any similar Walker / ParticleSet manipulations that might be lurking in the DMC sources. There are reports of the batched code paths being noisier than legacy, so perhaps there is some similar confusion.
Would like to know what the reports are especially those after Jan 2024.
Proposed changes
MCpopulation::saveWalkerConfigurations should only save data from internal walker configurations not particlesets. It affects DMC restart.
What type(s) of changes does this code introduce?
Does this introduce a breaking change?
What systems has this change been tested on?
epyc-server
Checklist