2.17.1-1
Add new NVLS algorithm for allreduce using NVLink SHARP (intra-node only).
Add new config options: cgaClusterSize, minCTAs, maxCTAs, netName.
Enable LL128 when we use PXN to close rings.
NVTX3 includes update.
Fix crash when one CollNet (SHARP) rail fails to initialize.
[ROCm/rccl commit: 5d3ab08b69]
Этот коммит содержится в:
@@ -74,6 +74,8 @@ void ncclDebugInit() {
|
||||
mask = NCCL_ALLOC;
|
||||
} else if (strcasecmp(subsys, "CALL") == 0) {
|
||||
mask = NCCL_CALL;
|
||||
} else if (strcasecmp(subsys, "NVLS") == 0) {
|
||||
mask = NCCL_NVLS;
|
||||
} else if (strcasecmp(subsys, "ALL") == 0) {
|
||||
mask = NCCL_ALL;
|
||||
}
|
||||
|
||||
Ссылка в новой задаче
Block a user