Merge remote-tracking branch 'nccl/master' into 2.8.3

[ROCm/rccl commit: c985358e11]
This commit is contained in:
Wenkai Du
2021-02-15 19:29:45 +00:00
23 changed files with 82 additions and 52 deletions
+4 -4
View File
@@ -58,7 +58,7 @@ static ncclResult_t findRevLink(struct ncclTopoNode* node1, struct ncclTopoNode*
return ncclSuccess;
}
}
WARN("Could not find rev link for %d/%d -> %d/%d\n", node1->type, node1->id, node2->type, node2->id);
WARN("Could not find rev link for %d/%ld -> %d/%ld", node1->type, node1->id, node2->type, node2->id);
return ncclInternalError;
}
@@ -191,7 +191,7 @@ static ncclResult_t getGpuIndex(struct ncclTopoSystem* system, int rank, int* in
return ncclSuccess;
}
}
WARN("Could not find gpu rank %d\n", rank);
WARN("Could not find gpu rank %d", rank);
return ncclInternalError;
}
@@ -202,7 +202,7 @@ static ncclResult_t getNetIndex(struct ncclTopoSystem* system, int64_t id, int*
return ncclSuccess;
}
}
WARN("Could not find net id %lx\n", id);
WARN("Could not find net id %lx", id);
return ncclInternalError;
}
@@ -1253,7 +1253,7 @@ done:
}
if (graph->nChannels == 0 && graph->collNet == 0) {
WARN("Could not find a path for pattern %d, falling back to simple order\n", graph->pattern);
WARN("Could not find a path for pattern %d, falling back to simple order", graph->pattern);
for (int i=0; i<ngpus; i++) graph->intra[i] = system->nodes[GPU].nodes[i].gpu.rank;
graph->inter[0] = graph->inter[1] = 0;
graph->speedIntra = graph->speedInter = 0.1;