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

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

Browse Source 
* 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
committed by GitHub
parent 9525c5b2ef
commit f3f3336468
6 changed files with 10 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/include/enqueue.h
+1 -1
View File
@@ -33,7 +33,7 @@ static inline size_t ncclFuncSendCount(ncclFunc_t func, int nRanks, size_t count
static inline size_t ncclFuncRecvCount(ncclFunc_t func, int nRanks, size_t count) {
return func == ncclFuncAllGather ? nRanks*count : count;
}
rccl_static_inline size_t ncclFuncMaxSendRecvCount(ncclFunc_t func, int nRanks, size_t count) {
rccl_static inline size_t ncclFuncMaxSendRecvCount(ncclFunc_t func, int nRanks, size_t count) {
return func == ncclFuncAllGather || func == ncclFuncReduceScatter ? nRanks*count : count;
}
src/include/rccl_common.h
-1
View File
@@ -78,5 +78,4 @@ ncclResult_t rcclGetAlgoInfo(struct ncclComm* comm, ncclFunc_t coll, uint64_t co
int* algo, int* protocol, int* maxChannels);
ncclResult_t rcclFuncMaxSendRecvCount(ncclFunc_t func, int nRanks, size_t count, size_t& maxCount);
#endif
src/rccl_wrap.cc
+1 -2
View File
@@ -23,7 +23,7 @@ THE SOFTWARE.
#include "rccl_common.h"
#include "comm.h"
#include "graph/topo.h"
#include "enqueue.h"
void rcclUpdateCollectiveProtocol(struct ncclComm* comm, size_t const& nBytes, struct ncclTaskColl* info) {
// Honor user input for protocol choice
static int userProtocolInput = -2;
@@ -72,7 +72,6 @@ void rcclUpdateCollectiveProtocol(struct ncclComm* comm, size_t const& nBytes, s
}
}
extern size_t ncclFuncMaxSendRecvCount(ncclFunc_t func, int nRanks, size_t count);
extern ncclResult_t getAlgoInfo(
struct ncclComm* comm, struct ncclTaskColl* task,
int collNetSupport, int nvlsSupport, int numPipeOps, ncclSimInfo_t* simInfo = NULL