apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

[BugFix]Dev fix custom ar unstable result (#4437)

Browse Source
This commit is contained in:
chen
2025-10-17 11:47:16 +08:00
committed by GitHub
parent 6160145f82
commit b134e6afe6
17 changed files with 25 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/distributed/custom_all_reduce/custom_all_reduce.py
+2 -2
View File
@@ -213,13 +213,13 @@ class CustomAllreduce:
stream_capturing = lib.cudaStreamIsCapturing(stream)
if stream_capturing.value == 1:
# 1 is cudaStreamCaptureStatusActive: The stream is capturing.
return self.all_reduce(input, input, registered=True)
return self.all_reduce(input, registered=True)
else:
# If warm up, mimic the allocation pattern since custom
# allreduce is out-of-place.
return paddle.empty_like(input)
else:
return self.all_reduce(input, input, registered=False)
return self.all_reduce(input, registered=False)
def clear_ipc_handles(self):
clear_ipc_handles(self._ptr)