54f37133b7
Mostly mechanical changes in preparation of extracting the Flang-RT "subproject" in #110217. This PR intends to only move pre-existing files to the new folder structure, with no behavioral change. Common files (headers, testing, cmake) shared by Flang-RT and Flang remain in `flang/`. Some cosmetic changes and files paths were necessary: * Relative paths to the new path for the source files and `add_subdirectory`. * Add the new location's include directory to `include_directories` * The unittest/Evaluate directory has unitests for flang-rt and Flang. A new `CMakeLists.txt` was introduced for the flang-rt tests. * Change the `#include` paths relative to the include directive * clang-format on the `#include` directives * Since the paths are part if the copyright header and include guards, a script was used to canonicalize those * `test/Runtime` and runtime tests in `test/Driver` are moved, but the lit.cfg.py mechanism to execute the will only be added in #110217.
72 lines
2.1 KiB
C++
72 lines
2.1 KiB
C++
//===-- lib/cuda/allocator.cpp ----------------------------------*- C++ -*-===//
|
|
//
|
|
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#include "flang/Runtime/CUDA/allocator.h"
|
|
#include "flang-rt/runtime/allocator-registry.h"
|
|
#include "flang-rt/runtime/derived.h"
|
|
#include "flang-rt/runtime/stat.h"
|
|
#include "flang-rt/runtime/terminator.h"
|
|
#include "flang-rt/runtime/type-info.h"
|
|
#include "flang/Common/ISO_Fortran_binding_wrapper.h"
|
|
#include "flang/Runtime/CUDA/common.h"
|
|
#include "flang/Support/Fortran.h"
|
|
|
|
#include "cuda_runtime.h"
|
|
|
|
namespace Fortran::runtime::cuda {
|
|
extern "C" {
|
|
|
|
void RTDEF(CUFRegisterAllocator)() {
|
|
allocatorRegistry.Register(
|
|
kPinnedAllocatorPos, {&CUFAllocPinned, CUFFreePinned});
|
|
allocatorRegistry.Register(
|
|
kDeviceAllocatorPos, {&CUFAllocDevice, CUFFreeDevice});
|
|
allocatorRegistry.Register(
|
|
kManagedAllocatorPos, {&CUFAllocManaged, CUFFreeManaged});
|
|
allocatorRegistry.Register(
|
|
kUnifiedAllocatorPos, {&CUFAllocUnified, CUFFreeUnified});
|
|
}
|
|
}
|
|
|
|
void *CUFAllocPinned(std::size_t sizeInBytes) {
|
|
void *p;
|
|
CUDA_REPORT_IF_ERROR(cudaMallocHost((void **)&p, sizeInBytes));
|
|
return p;
|
|
}
|
|
|
|
void CUFFreePinned(void *p) { CUDA_REPORT_IF_ERROR(cudaFreeHost(p)); }
|
|
|
|
void *CUFAllocDevice(std::size_t sizeInBytes) {
|
|
void *p;
|
|
CUDA_REPORT_IF_ERROR(cudaMalloc(&p, sizeInBytes));
|
|
return p;
|
|
}
|
|
|
|
void CUFFreeDevice(void *p) { CUDA_REPORT_IF_ERROR(cudaFree(p)); }
|
|
|
|
void *CUFAllocManaged(std::size_t sizeInBytes) {
|
|
void *p;
|
|
CUDA_REPORT_IF_ERROR(
|
|
cudaMallocManaged((void **)&p, sizeInBytes, cudaMemAttachGlobal));
|
|
return reinterpret_cast<void *>(p);
|
|
}
|
|
|
|
void CUFFreeManaged(void *p) { CUDA_REPORT_IF_ERROR(cudaFree(p)); }
|
|
|
|
void *CUFAllocUnified(std::size_t sizeInBytes) {
|
|
// Call alloc managed for the time being.
|
|
return CUFAllocManaged(sizeInBytes);
|
|
}
|
|
|
|
void CUFFreeUnified(void *p) {
|
|
// Call free managed for the time being.
|
|
CUFFreeManaged(p);
|
|
}
|
|
|
|
} // namespace Fortran::runtime::cuda
|