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: 36343be84f]
This commit is contained in:
BertanDogancay
2025-01-23 11:48:18 -06:00
committed by Corey Derochie
parent e3a29f5eab 5ca1b6c160
commit 1b000665df
122 changed files with 9009 additions and 4702 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rccl/src/include/graph.h
+5 -2
View File
@@ -30,6 +30,7 @@ ncclResult_t ncclTopoTrimSystem(struct ncclTopoSystem* system, struct ncclComm*
ncclResult_t ncclTopoComputeP2pChannels(struct ncclComm* comm);
ncclResult_t ncclTopoGetNvbGpus(struct ncclTopoSystem* system, int rank, int* nranks, int** ranks);
int ncclTopoPathAllNVLink(struct ncclTopoSystem* system);
ncclResult_t ncclTopoComputeCommCPU(struct ncclComm* comm);
// Query topology
ncclResult_t ncclTopoGetNetDev(struct ncclComm* comm, int rank, struct ncclTopoGraph* graph, int channelId, int peerRank, int64_t* id, int* dev, int* proxyRank);
@@ -50,9 +51,11 @@ ncclResult_t ncclTopoGetCpuAffinity(struct ncclTopoSystem* system, int rank, cpu
#define NCCL_TOPO_CPU_ARCH_X86 1
#define NCCL_TOPO_CPU_ARCH_POWER 2
#define NCCL_TOPO_CPU_ARCH_ARM 3
#define NCCL_TOPO_CPU_ARCH_MIXED 4
#define NCCL_TOPO_CPU_VENDOR_INTEL 1
#define NCCL_TOPO_CPU_VENDOR_AMD 2
#define NCCL_TOPO_CPU_VENDOR_ZHAOXIN 3
#define NCCL_TOPO_CPU_VENDOR_MIXED 4
#define NCCL_TOPO_CPU_TYPE_BDW 1
#define NCCL_TOPO_CPU_TYPE_SKL 2
#define NCCL_TOPO_CPU_TYPE_ZEN 3
@@ -76,6 +79,7 @@ ncclResult_t ncclTopoSearchInit(struct ncclTopoSystem* system);
#define NCCL_TOPO_PATTERN_TREE 3 // All NIC traffic going to/from the same GPU
#define NCCL_TOPO_PATTERN_RING 4 // Ring
#define NCCL_TOPO_PATTERN_NVLS 5 // NVLS+SHARP and NVLS+Tree
#define NCCL_TOPO_PATTERN_COLLNET_DIRECT 6 // Collnet Direct
struct ncclTopoGraph {
// Input / output
int id; // ring : 0, tree : 1, collnet : 2
@@ -123,7 +127,6 @@ ncclResult_t ncclTopoPostset(struct ncclComm* comm, int* firstRanks, int* treePa
ncclResult_t ncclTreeBasePostset(struct ncclComm* comm, struct ncclTopoGraph* treeGraph);
ncclResult_t ncclTopoTuneModel(struct ncclComm* comm, int minCompCap, int maxCompCap, struct ncclTopoGraph** graphs);
#include "info.h"
ncclResult_t ncclTopoGetAlgoTime(struct ncclInfo* info, int algorithm, int protocol, int numPipeOps, float* time, bool* backup = NULL);
ncclResult_t ncclTopoGetAlgoTime(struct ncclComm* comm, int coll, int algorithm, int protocol, size_t nBytes, int numPipeOps, float* time, bool* backup=nullptr);
#endif