Sylvain Jeaugey
20da390b96
Add support for CUDA graphs.
Fuse BCM Gen4 switches to avoid suboptimal performance on some platforms. Issue #439.
Fix bootstrap issue caused by connection reordering.
Fix CPU locking block.
Improve CollNet algorithm.
Improve performance on DGX A100 for communicators with only one GPU per node.
[ROCm/rccl commit: a46ea10583]
72 рядки
2.1 KiB
C++
72 рядки
2.1 KiB
C++
/*************************************************************************
|
|
* Copyright (c) 2019-2021, NVIDIA CORPORATION. All rights reserved.
|
|
*
|
|
* See LICENSE.txt for license information
|
|
************************************************************************/
|
|
|
|
#ifndef NCCL_ALLOC_H_
|
|
#define NCCL_ALLOC_H_
|
|
|
|
#include "nccl.h"
|
|
#include "checks.h"
|
|
#include "align.h"
|
|
#include <sys/mman.h>
|
|
|
|
template <typename T>
|
|
static ncclResult_t ncclCudaHostCalloc(T** ptr, size_t nelem) {
|
|
CUDACHECK(cudaHostAlloc(ptr, nelem*sizeof(T), cudaHostAllocMapped));
|
|
memset(*ptr, 0, nelem*sizeof(T));
|
|
return ncclSuccess;
|
|
}
|
|
|
|
static inline ncclResult_t ncclCudaHostFree(void* ptr) {
|
|
CUDACHECK(cudaFreeHost(ptr));
|
|
return ncclSuccess;
|
|
}
|
|
|
|
template <typename T>
|
|
static ncclResult_t ncclCalloc(T** ptr, size_t nelem) {
|
|
void* p = malloc(nelem*sizeof(T));
|
|
if (p == NULL) {
|
|
WARN("Failed to malloc %ld bytes", nelem*sizeof(T));
|
|
return ncclSystemError;
|
|
}
|
|
memset(p, 0, nelem*sizeof(T));
|
|
*ptr = (T*)p;
|
|
return ncclSuccess;
|
|
}
|
|
|
|
template <typename T>
|
|
static ncclResult_t ncclCudaCalloc(T** ptr, size_t nelem) {
|
|
// Need async stream for P2P pre-connect + CUDA Graph
|
|
cudaStream_t stream;
|
|
CUDACHECK(cudaStreamCreateWithFlags(&stream, cudaStreamNonBlocking));
|
|
CUDACHECK(cudaMalloc(ptr, nelem*sizeof(T)));
|
|
CUDACHECK(cudaMemsetAsync(*ptr, 0, nelem*sizeof(T), stream));
|
|
CUDACHECK(cudaStreamSynchronize(stream));
|
|
CUDACHECK(cudaStreamDestroy(stream));
|
|
return ncclSuccess;
|
|
}
|
|
|
|
template <typename T>
|
|
static ncclResult_t ncclCudaMemcpy(T* dst, T* src, size_t nelem) {
|
|
CUDACHECK(cudaMemcpy(dst, src, nelem*sizeof(T), cudaMemcpyDefault));
|
|
return ncclSuccess;
|
|
}
|
|
|
|
// Allocate memory to be potentially ibv_reg_mr'd. This needs to be
|
|
// allocated on separate pages as those pages will be marked DONTFORK
|
|
// and if they are shared, that could cause a crash in a child process
|
|
static ncclResult_t ncclIbMalloc(void** ptr, size_t size) {
|
|
size_t page_size = sysconf(_SC_PAGESIZE);
|
|
void* p;
|
|
int size_aligned = ROUNDUP(size, page_size);
|
|
int ret = posix_memalign(&p, page_size, size_aligned);
|
|
if (ret != 0) return ncclSystemError;
|
|
memset(p, 0, size);
|
|
*ptr = p;
|
|
return ncclSuccess;
|
|
}
|
|
|
|
#endif
|