Sylvain Jeaugey
19ab67d172
Optimize CUDA graph launch; avoid launching a CPU callback for intra-node operations. Simplify kernel common code to improve the latency of send/recv operations. Strengthen CUDA streams semantics. Change NET API to v6, to add dmabuf support. Add ncclGetLastError() function. Add ncclRemoteError code and use it for remote network errors. Support the use of a different NCCL_NET parameter per communicator. Add support for SHM and P2P transfers using cudaMemcpy.
48 خطوط
1.5 KiB
C++
48 خطوط
1.5 KiB
C++
/*************************************************************************
|
|
* Copyright (c) 2015-2022, NVIDIA CORPORATION. All rights reserved.
|
|
*
|
|
* See LICENSE.txt for license information
|
|
************************************************************************/
|
|
|
|
#ifndef NCCL_DEBUG_H_
|
|
#define NCCL_DEBUG_H_
|
|
|
|
#include "nccl_net.h"
|
|
#include <stdio.h>
|
|
#include <chrono>
|
|
#include <type_traits>
|
|
|
|
#include <limits.h>
|
|
#include <string.h>
|
|
#include <pthread.h>
|
|
|
|
// Conform to pthread and NVTX standard
|
|
#define NCCL_THREAD_NAMELEN 16
|
|
|
|
extern int ncclDebugLevel;
|
|
extern uint64_t ncclDebugMask;
|
|
extern pthread_mutex_t ncclDebugLock;
|
|
extern FILE *ncclDebugFile;
|
|
extern ncclResult_t getHostName(char* hostname, int maxlen, const char delim);
|
|
|
|
void ncclDebugLog(ncclDebugLogLevel level, unsigned long flags, const char *filefunc, int line, const char *fmt, ...) __attribute__ ((format (printf, 5, 6)));
|
|
|
|
// Let code temporarily downgrade WARN into INFO
|
|
extern thread_local int ncclDebugNoWarn;
|
|
extern char ncclLastError[];
|
|
|
|
#define WARN(...) ncclDebugLog(NCCL_LOG_WARN, NCCL_ALL, __FILE__, __LINE__, __VA_ARGS__)
|
|
#define INFO(FLAGS, ...) ncclDebugLog(NCCL_LOG_INFO, (FLAGS), __func__, __LINE__, __VA_ARGS__)
|
|
#define TRACE_CALL(...) ncclDebugLog(NCCL_LOG_TRACE, NCCL_CALL, __func__, __LINE__, __VA_ARGS__)
|
|
|
|
#ifdef ENABLE_TRACE
|
|
#define TRACE(FLAGS, ...) ncclDebugLog(NCCL_LOG_TRACE, (FLAGS), __func__, __LINE__, __VA_ARGS__)
|
|
extern std::chrono::steady_clock::time_point ncclEpoch;
|
|
#else
|
|
#define TRACE(...)
|
|
#endif
|
|
|
|
void ncclSetThreadName(pthread_t thread, const char *fmt, ...);
|
|
|
|
#endif
|