AI/rocm-systems
2
0
Forkuj 0
Kod Zgłoszenia Oczekujące zmiany Actions Packages Projekty Wydania Wiki Aktywność

Fix topo explorer's compatibility with NCCL 2.24 (#1671)

Przeglądaj źródła 
* Fix build issues

* Fix failure to find path remote rank
This commit is contained in:
Mustafa Abduljabbar
2025-05-05 15:26:29 -04:00
zatwierdzone przez GitHub
rodzic 9525c5b2ef
commit f3f3336468
6 zmienionych plików z 10 dodań i 12 usunięć
Ściągnij plik aktualizacji Ściągnij plik porównania Expand all files Collapse all files
tools/topo_expl/model.cpp
+5 -5
Wyświetl plik
@@ -92,7 +92,7 @@ ncclResult_t p2pCanConnect(int* ret, struct ncclComm* comm, struct ncclTopoGrap
// Check topology / p2p level.
int intermediateRank;
NCCLCHECK(ncclTopoCheckP2p(comm->topo, info1->rank, info2->rank, ret, NULL, &intermediateRank));
NCCLCHECK(ncclTopoCheckP2p(comm, comm->topo, info1->rank, info2->rank, ret, NULL, &intermediateRank));
if (*ret == 0) return ncclSuccess;
if (intermediateRank != -1) {
if (useMemcpy) *ret = 0;
@@ -214,7 +214,7 @@ ncclResult_t netSendSetup(struct ncclComm* comm, struct ncclTopoGraph* graph, st
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));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->rank, 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,
@@ -242,7 +242,7 @@ ncclResult_t netRecvSetup(struct ncclComm* comm, struct ncclTopoGraph* graph, st
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));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->rank, 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" : "");
@@ -266,7 +266,7 @@ ncclResult_t collNetSendSetup(struct ncclComm* comm, struct ncclTopoGraph* graph
int netDev, useGdr = 0, proxy;
NCCLCHECK(ncclTopoGetNetDev(comm, myInfo->rank, graph, channelId, peerInfo->rank, NULL, &netDev, &proxy));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->busId, netDev, 1, &useGdr));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->rank, netDev, 1, &useGdr));
INFO(NCCL_INIT|NCCL_NET,"Coll %02d : %d [send] via COLLNET/%s/%d%s", channelId, myInfo->rank, "SHARP", netDev, useGdr ? "/GDRDMA" : "");
return ncclSuccess;
@@ -276,7 +276,7 @@ ncclResult_t collNetRecvSetup(struct ncclComm* comm, struct ncclTopoGraph* graph
int netDev, useGdr = 0, proxy;
NCCLCHECK(ncclTopoGetNetDev(comm, myInfo->rank, graph, channelId, peerInfo->rank, NULL, &netDev, &proxy));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->busId, netDev, 0, &useGdr));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->rank, netDev, 0, &useGdr));
INFO(NCCL_INIT|NCCL_NET,"Coll %02d : %d [receive] via COLLNET/%s/%d%s", channelId, myInfo->rank, "SHARP", netDev, useGdr ? "/GDRDMA" : "");
return ncclSuccess;