AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

2.15.1-1

Browse Source 
Add support for H100 (sm90).
Make sure NCCL kernel honor user stream priorities.


[ROCm/rccl commit: da8152e57a]
This commit is contained in:
Sylvain Jeaugey
2022-09-27 02:31:13 -07:00
parent 37ccaf1f82
commit b4bac0d15a
22 changed files with 303 additions and 76 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rccl/src/graph/search.cc
+1 -1
View File
@@ -727,7 +727,7 @@ ncclResult_t ncclTopoGetXmlFromGraphs(int ngraphs, struct ncclTopoGraph** graphs
}
float speedArrayIntra[] = { 44.0, 30.0, 22.0, 18.0, 15.0, 12.0, 10.0, 9.0, 7.0, 6.0, 5.0, 4.0, 3.0 };
float speedArrayInter[] = { 48.0, 30.0, 24.0, 22.0, 18.0, 15.0, 12.0, 10.0, 9.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.4, 1.2, 0.24, 0.12 };
float speedArrayInter[] = { 48.0, 30.0, 28.0, 24.0, 22.0, 18.0, 15.0, 12.0, 10.0, 9.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.4, 1.2, 0.24, 0.12 };
#define NSPEEDSINTRA (sizeof(speedArrayIntra)/sizeof(float))
#define NSPEEDSINTER (sizeof(speedArrayInter)/sizeof(float))