frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00
Code Issues Packages Projects Releases Wiki Activity

revert PTD's change that leads to signature mismatch of printNcclCommProxyTrace (#146453)

Browse Source 
Summary: D68801098 introduced this function signature mismatch issue for printNcclCommProxyTrace. Revert it so that trunk build can pass.

Test Plan:
With the change, build of APS model using rcclexp can now pass:
`sh scripts/ltian/run_jobs/fb_fm_v2/run_fb_fm_v2_job.sh -h T20_GTT_MI300X -n 16 -b 1024 -t [2024-12-06] -d ai_infra_ngs -e ai_infra_training_rnd_tc -x 0`

Reviewed By: c-p-i-o

Differential Revision: D69149588

Pull Request resolved: https://github.com/pytorch/pytorch/pull/146453
Approved by: https://github.com/c-p-i-o
This commit is contained in:
Dingming Wu
2025-02-07 22:43:52 +00:00
committed by PyTorch MergeBot
parent 103c8b44bc
commit fa34128435
2 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
torch/csrc/distributed/c10d/NCCLUtils.cpp
View File

@ -549,7 +549,7 @@ std::string getNcclErrorDetailStr(
// Dump proxyTrace log to stdout // Dump proxyTrace log to stdout
void printNcclCommProxyTrace( void printNcclCommProxyTrace(
std::string& dumpReason, const std::string& dumpReason,
const std::unordered_map<std::string, std::string>& dumpMap) { const std::unordered_map<std::string, std::string>& dumpMap) {
LOG(INFO) << "Dumping nccl comm trace, reason: " << dumpReason; LOG(INFO) << "Dumping nccl comm trace, reason: " << dumpReason;
for (auto& [key, value] : dumpMap) { for (auto& [key, value] : dumpMap) {

2
torch/csrc/distributed/c10d/NCCLUtils.hpp
View File

@ -349,7 +349,7 @@ struct ncclRedOpRAII {
}; };
void printNcclCommProxyTrace( void printNcclCommProxyTrace(
std::string dumpReason, const std::string& dumpReason,
const std::unordered_map<std::string, std::string>& dumpMap); const std::unordered_map<std::string, std::string>& dumpMap);
} // namespace c10d } // namespace c10d