AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge remote-tracking branch 'nccl/master' into develop

Browse Source
This commit is contained in:
Wenkai Du
2020-06-08 20:45:19 -07:00
parent 87db65f22d 5949d96f36
commit 26a0fd2517
56 changed files with 2054 additions and 1264 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/include/alloc.h
+5 -5
View File
@@ -1,5 +1,5 @@
/*************************************************************************
* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* Copyright (c) 2019-2020, NVIDIA CORPORATION. All rights reserved.
* Modifications Copyright (c) 2019-2020 Advanced Micro Devices, Inc. All rights reserved.
*
* See LICENSE.txt for license information
@@ -13,10 +13,10 @@
#include "align.h"
#include <sys/mman.h>
static inline ncclResult_t ncclCudaHostAlloc(void** ptr, void** devPtr, size_t size) {
CUDACHECK(hipHostMalloc(ptr, size, hipHostMallocMapped));
memset(*ptr, 0, size);
*devPtr = *ptr;
template <typename T>
static ncclResult_t ncclCudaHostCalloc(T** ptr, size_t nelem) {
CUDACHECK(hipHostMalloc(ptr, nelem*sizeof(T), hipHostMallocMapped));
memset(*ptr, 0, nelem*sizeof(T));
return ncclSuccess;
}