Open cpviolator opened 1 year ago
Nothing controversial in this commit. One thought on the design used here: would it be a better fit to push
and pop
gauge / momentum fields. E.g., instead of backing up the resident fields and zeroing them, and the later restoring them, we just push
a new field (which pushes the resident field onto the proverbial stack) and when we're done with this temporary, we simply pop
it, freeing the temporary and restoring the prior resident field off of the stack.
Just a thought.
This WIP add functionality to allow for offload in the FGI RHMC routines in https://github.com/callat-qcd/milc_qcd/tree/feature/nested-force-gradient-integrator. It simply creates copy gauge and momentum fields consistent with MILC's QUDA offload sequences.