mirror of
https://github.com/pytorch/pytorch.git
synced 2025-10-21 13:44:15 +08:00
1ad0048b64
Headers under torch/csrc/distributed may be referened with relative path, e.g., "<c10d/...>". However, relative path cannot be gracefully handled by Meta internal build when the NCCL PG is hipified to support AMD/RCCL because the "hipified" header files are generated in other directories. Moreover, using absolute path for header inclusion is the state-of-the-art in most components in Pytorch. Thus, this patch refactors all header paths in torch/csrc/distributed to be absolute. See D39835774 for more details about Meta internal complication. **How to test**: commit 9e5d199 removes -I./torch/csrc/distributed in compile options. Thus use it to verify we don't miss any relative path use of torch/csrc/distributed headers. Pull Request resolved: https://github.com/pytorch/pytorch/pull/85780 Approved by: https://github.com/kumpera, https://github.com/huydhn
59 lines
2.2 KiB
C++
59 lines
2.2 KiB
C++
#pragma once
|
|
|
|
#ifdef USE_C10D_UCC
|
|
|
|
#include <torch/csrc/distributed/c10d/UCCUtils.hpp>
|
|
|
|
namespace c10d {
|
|
|
|
#define RECORD_COMMS_TRACE( \
|
|
_comms_tracer, _work, _opType, _rank, _comm_size, _inTensors, _outTensors) \
|
|
do { \
|
|
if (torch_ucc_config.enable_comms_logger) { \
|
|
_comms_tracer->recordComms( \
|
|
opTypeToString(_opType), \
|
|
(uintptr_t)_work.get(), \
|
|
_rank, \
|
|
_comm_size, \
|
|
_inTensors, \
|
|
_outTensors); \
|
|
} \
|
|
} while (0)
|
|
|
|
// interfaces to collect communication traces
|
|
class TORCH_API CommTraceLogger : public torch::CustomClassHolder {
|
|
private:
|
|
std::vector<std::string> comms_trace_;
|
|
std::vector<std::string> curBlocks_; /* unused */
|
|
std::vector<int64_t> curOutSplitSizes_;
|
|
std::vector<int64_t> curInSplitSizes_;
|
|
int curRoot_ = -1;
|
|
unsigned long seqnum = 0;
|
|
|
|
public:
|
|
void setCurBlock(const std::string& name); /* unused */
|
|
void popBlock(); /* unused */
|
|
// record root info if applicable, e.g., broadcast, gather, scatter
|
|
void recordOptionalInfo(int root = -1);
|
|
// record input/output splits of Alltoallv
|
|
void recordOptionalInfo(
|
|
const std::vector<int64_t>& outputSplitSizes = {},
|
|
const std::vector<int64_t>& inputSplitSizes = {});
|
|
// record essential comms information
|
|
void recordComms(
|
|
const std::string& collName,
|
|
const uintptr_t workReq = 0,
|
|
const int rank = -1,
|
|
const int world_size = -1,
|
|
const std::vector<at::Tensor>& inputTensors = {},
|
|
const std::vector<at::Tensor>& outputTensor = {});
|
|
// return collected comms traces
|
|
std::vector<std::string>& getCommsTrace() {
|
|
return comms_trace_;
|
|
}
|
|
};
|
|
|
|
} // namespace c10d
|
|
|
|
#endif // USE_C10D_UCC
|