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/src/include/graph.h
+7 -7
View File
@@ -32,12 +32,12 @@ ncclResult_t ncclTopoGetNvbGpus(struct ncclTopoSystem* system, int rank, int* nr
int ncclTopoPathAllNVLink(struct ncclTopoSystem* system);
// Query topology
ncclResult_t ncclTopoGetNetDev(struct ncclComm* comm, int rank, struct ncclTopoGraph* graph, int channelId, int peerRank, int* net, int* proxyRank);
ncclResult_t ncclTopoGetNetDev(struct ncclComm* comm, int rank, struct ncclTopoGraph* graph, int channelId, int peerRank, int64_t* id, int* dev, int* proxyRank);
ncclResult_t ncclTopoCheckP2p(struct ncclTopoSystem* system, int64_t id1, int64_t id2, int* p2p, int *read, int* intermediateRank);
ncclResult_t ncclTopoCheckMNNVL(struct ncclTopoSystem* system, struct ncclPeerInfo* info1, struct ncclPeerInfo* info2, int* ret);
ncclResult_t ncclTopoCheckGdr(struct ncclTopoSystem* topo, int64_t busId, int netDev, int read, int* useGdr);
ncclResult_t ncclTopoCheckGdr(struct ncclTopoSystem* topo, int64_t busId, int64_t netId, int read, int* useGdr);
#define MAX_XGMI_INTER_GPUS 4
ncclResult_t ncclTopoGetIntraNetDev(struct ncclTopoSystem* system, int rank, struct ncclTopoGraph* graph, int channelId, int type, int* dev);
ncclResult_t ncclTopoGetIntraNetDev(struct ncclTopoSystem* system, int rank, struct ncclTopoGraph* graph, int channelId, int type, int64_t* id, int* dev);
ncclResult_t ncclTopoGetLinkType(struct ncclTopoSystem* system, int cudaDev1, int cudaDev2, bool* isXGMI, int maxInter=MAX_XGMI_INTER_GPUS, int nInter=0, int *inter=nullptr);
ncclResult_t ncclTopoNeedFlush(struct ncclTopoSystem* system, int64_t busId, int* flush);
ncclResult_t ncclTopoCheckNet(struct ncclTopoSystem* system, int64_t id1, int64_t id2, int* net);
@@ -62,8 +62,8 @@ ncclResult_t ncclTopoCpuType(struct ncclTopoSystem* system, int* arch, int* vend
ncclResult_t ncclTopoGetGpuCount(struct ncclTopoSystem* system, int* count);
ncclResult_t ncclTopoGetNetCount(struct ncclTopoSystem* system, int* count);
ncclResult_t ncclTopoGetNvsCount(struct ncclTopoSystem* system, int* count);
ncclResult_t ncclTopoGetLocalNet(struct ncclTopoSystem* system, int rank, int channelId, int* id);
ncclResult_t ncclTopoGetLocalGpu(struct ncclTopoSystem* system, int net, int* gpuIndex);
ncclResult_t ncclTopoGetLocalNet(struct ncclTopoSystem* system, int rank, int channelId, int64_t* id, int* dev);
ncclResult_t ncclTopoGetLocalGpu(struct ncclTopoSystem* system, int64_t netId, int* gpuIndex);
ncclResult_t getLocalNetCountByBw(struct ncclTopoSystem* system, int gpu, int *count);
#define NCCL_TOPO_MAX_NODES 64
@@ -94,7 +94,7 @@ struct ncclTopoGraph {
int sameChannels;
int nHops;
int intra[MAXCHANNELS*NCCL_TOPO_MAX_NODES];
int inter[MAXCHANNELS*2];
int64_t inter[MAXCHANNELS*2];
int nIntraChannels;
int intraNets[MAXCHANNELS*NCCL_TOPO_MAX_NODES*2];
char treeBase[NCCL_TOPO_MAX_NODES][NCCL_TOPO_MAX_NODES*4];
@@ -119,7 +119,7 @@ struct ncclTopoRanks {
ncclResult_t ncclTopoPreset(struct ncclComm* comm, struct ncclTopoGraph** graphs, struct ncclTopoRanks* topoRanks);
ncclResult_t ncclTopoPostset(struct ncclComm* comm, int* firstRanks, int* treePatterns,
struct ncclTopoRanks** allTopoRanks, int* rings, struct ncclTopoGraph** graphs, int nc);
struct ncclTopoRanks** allTopoRanks, int* rings, struct ncclTopoGraph** graphs, struct ncclComm* parent, int nc);
ncclResult_t ncclTreeBasePostset(struct ncclComm* comm, struct ncclTopoGraph* treeGraph);
ncclResult_t ncclTopoTuneModel(struct ncclComm* comm, int minCompCap, int maxCompCap, struct ncclTopoGraph** graphs);