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

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

Browse Source 
[ROCm/rccl commit: 84081064a0]
This commit is contained in:
BertanDogancay
2024-10-02 09:29:22 -05:00
parent 74aa13afbe 792c33598f
commit 9059445acb
54 changed files with 2158 additions and 964 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rccl/tools/topo_expl/model.cpp
+8 -6
View File
@@ -211,9 +211,10 @@ ncclResult_t netSendSetup(struct ncclComm* comm, struct ncclTopoGraph* graph, st
req.netDev = -1;
int proxyRank = myInfo->rank;
if (connIndex == NCCL_CONN_IDX_P2P_NET) NCCLCHECK(ncclTopoGetIntraNetDev(comm->topo, myInfo->rank, graph, channelId, 1, &req.netDev));
if (req.netDev < 0) NCCLCHECK(ncclTopoGetNetDev(comm, myInfo->rank, graph, channelId, peerInfo->rank, &req.netDev, &proxyRank));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->busId, req.netDev, 1, &req.useGdr));
int64_t netId;
if (connIndex == NCCL_CONN_IDX_P2P_NET) NCCLCHECK(ncclTopoGetIntraNetDev(comm->topo, myInfo->rank, graph, channelId, 1, &netId, &req.netDev));
if (req.netDev < 0) NCCLCHECK(ncclTopoGetNetDev(comm, myInfo->rank, graph, channelId, peerInfo->rank, &netId, &req.netDev, &proxyRank));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->busId, netId, 1, &req.useGdr));
if (proxyRank == myInfo->rank) {
INFO(NCCL_INIT|NCCL_NET,"Channel %02d/%d : %d[%lx] -> %d[%lx] [send] via NET/%s/%d%s%s", channelId, connIndex, myInfo->rank, myInfo->busId, peerInfo->rank, peerInfo->busId, comm->ncclNet->name, req.netDev,
@@ -238,9 +239,10 @@ ncclResult_t netRecvSetup(struct ncclComm* comm, struct ncclTopoGraph* graph, st
// Use myInfo->rank as the receiver uses its own NIC
int proxyRank = myInfo->rank;
if (connIndex == NCCL_CONN_IDX_P2P_NET) NCCLCHECK(ncclTopoGetIntraNetDev(comm->topo, myInfo->rank, graph, channelId, 0, &req.netDev));
if (req.netDev < 0) NCCLCHECK(ncclTopoGetNetDev(comm, myInfo->rank, graph, channelId, myInfo->rank, &req.netDev, &proxyRank));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->busId, req.netDev, 0, &req.useGdr));
int64_t netId;
if (connIndex == NCCL_CONN_IDX_P2P_NET) NCCLCHECK(ncclTopoGetIntraNetDev(comm->topo, myInfo->rank, graph, channelId, 0, &netId, &req.netDev));
if (req.netDev < 0) NCCLCHECK(ncclTopoGetNetDev(comm, myInfo->rank, graph, channelId, myInfo->rank, &netId, &req.netDev, &proxyRank));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->busId, netId, 0, &req.useGdr));
INFO(NCCL_INIT|NCCL_NET,"Channel %02d/%d : %d[%lx] -> %d[%lx] [receive] via NET/%s/%d%s%s", channelId, connIndex, peerInfo->rank, peerInfo->busId, myInfo->rank, myInfo->busId, comm->ncclNet->name, req.netDev,
req.useGdr ? "/GDRDMA" : "", req.shared ? "/Shared" : "");
projects/rccl/tools/topo_expl/utils.cpp
+2 -2
View File
@@ -342,8 +342,8 @@ static ncclResult_t selectTransport(struct ncclComm* comm, struct ncclTopoGraph*
// handle intra-node network connections
int n1 = -1, n2 = -1;
if (connIndex == NCCL_CONN_IDX_P2P_NET) {
NCCLCHECK(ncclTopoGetIntraNetDev(comm->topo, comm->rank, graph, channelId, (type == 1) ? 1 : 0, &n1));
NCCLCHECK(ncclTopoGetIntraNetDev(comm->topo, peer, graph, channelId, (type == 1) ? 0 : 1, &n2));
NCCLCHECK(ncclTopoGetIntraNetDev(comm->topo, comm->rank, graph, channelId, (type == 1) ? 1 : 0, nullptr, &n1));
NCCLCHECK(ncclTopoGetIntraNetDev(comm->topo, peer, graph, channelId, (type == 1) ? 0 : 1, nullptr, &n2));
}
bool xgmi;
NCCLCHECK(ncclTopoGetLinkType(comm->topo, myInfo->cudaDev, peerInfo->cudaDev, &xgmi));