You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fix#2619.
The GPU implementation in this PR is usually faster than the CPU in one
thread (i.e., not using the feature implemented in #1624). Still, it
needs parallelism in the batch dimension, which is blocked by #2618,
regarding building the neighbor list. The GPU utilization is less than
10% for the water system. It should be improved when #2618 makes
progress.
---------
Signed-off-by: Jinzhe Zeng <[email protected]>
Summary
Currently, the neighbor stat is conducted in CPUs. It is slower than that on GPUs, but we do not have GPU implementation.
Detailed Description
The current neighbor stat code is
deepmd-kit/deepmd/utils/neighbor_stat.py
Lines 75 to 82 in 92ca097
Ideally, parallelism in the dimension of the frames is also required to handle many structures.
Further Information, Files, and Links
No response
The text was updated successfully, but these errors were encountered: