Fix bug #307 : wrong NIC selection on the reduction tree.

The reduction tree (tree up) was inverting the NICs to use,
causing performance issue in cases where we are using different
NICs on a given channel.


[ROCm/rccl commit: b5b6c6acdd]
This commit is contained in:
Sylvain Jeaugey
2020-04-09 17:14:07 -07:00
parent b996c2ca00
commit 627e1a06d0
4 ha cambiato i file con 10 aggiunte e 7 eliminazioni
+1 -1
Vedi File
@@ -27,7 +27,7 @@ void ncclTopoFree(struct ncclTopoSystem* system);
ncclResult_t ncclTopoTrimSystem(struct ncclTopoSystem* system, struct ncclComm* comm);
// Query topology
ncclResult_t ncclTopoGetNetDev(struct ncclTopoGraph* graph, int dir, int channelId, int* net);
ncclResult_t ncclTopoGetNetDev(struct ncclTopoSystem* system, struct ncclTopoGraph* graph, int rank, int channelId, int* net);
ncclResult_t ncclTopoCheckP2p(struct ncclTopoSystem* system, int64_t id1, int64_t id2, int* p2p);
ncclResult_t ncclTopoCheckGdr(struct ncclTopoSystem* topo, int64_t busId, int netDev, int read, int* useGdr);