[rocm_regression] Return errors when HSA_NO_SCRATCH_RECLAIM=1 even for rocm>=6.4.0 (#1867)

* [rocm_regression] Return errors when HSA_NO_SCRATCH_RECLAIM=1 even for rocm >= 6.4.0
* [rocm_regression] Check firmware version
* [rocm_regression] Resolve review comments
* [rocm_regression] Move hsa env checking into init once func
* [rocm_regression] Prevent hot fix version in firmware
* [rocm_regression] Improve unit tests
Dieser Commit ist enthalten in:
ycui1984
2025-08-29 09:18:23 -07:00
committet von GitHub
Ursprung 9afc15625f
Commit 361d596229
4 geänderte Dateien mit 124 neuen und 9 gelöschten Zeilen
+2
Datei anzeigen
@@ -91,4 +91,6 @@ void rcclSetPxn(struct ncclComm* comm, int& rcclPxnDisable);
void rcclSetP2pNetChunkSize(struct ncclComm* comm, int& rcclP2pNetChunkSize);
ncclResult_t rcclFuncMaxSendRecvCount(ncclFunc_t func, int nRanks, size_t count, size_t& maxCount);
ncclResult_t commSetUnrollFactor(struct ncclComm* comm);
bool validHsaScratchEnvSetting(const char*hsaScratchEnv, int hipRuntimeVersion, int firmwareVersion, const char* archName);
int parseFirmwareVersion(const char* command);
#endif