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]
Este cometimento está contido em:
@@ -315,7 +315,7 @@ static ncclResult_t groupLaunch(struct ncclAsyncJob *job_) {
|
||||
ret = ncclSystemError;
|
||||
}
|
||||
job->state = ncclGroupJobJoined;
|
||||
if (job->result != ncclSuccess) {
|
||||
if (job->result != ncclSuccess && ret == ncclSuccess) {
|
||||
ret = job->result;
|
||||
errorJobAbortFlag = true;
|
||||
}
|
||||
@@ -326,7 +326,6 @@ static ncclResult_t groupLaunch(struct ncclAsyncJob *job_) {
|
||||
|
||||
if (*groupAbortFlag == true || errorJobAbortFlag == true) {
|
||||
*job->abortFlag = 1;
|
||||
ret = ncclInternalError;
|
||||
}
|
||||
|
||||
job = job->next;
|
||||
|
||||
Criar uma nova questão referindo esta
Bloquear um utilizador