cpu_info.h 1.7 KB
Newer Older
1
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved.
L
liaogang 已提交
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License. */

#pragma once

#include <stddef.h>

D
dzhwinter 已提交
19 20 21 22 23 24 25 26 27 28 29 30
#ifdef _WIN32
#if defined(__AVX2__)
#include <immintrin.h>  //avx2
#elif defined(__AVX__)
#include <intrin.h>  //avx
#endif               // AVX
#else                // WIN32
#ifdef __AVX__
#include <immintrin.h>
#endif
#endif  // WIN32

L
liaogang 已提交
31 32 33
namespace paddle {
namespace platform {

L
liaogang 已提交
34 35 36
//! Get the maximum allocation size for a machine.
size_t CpuMaxAllocSize();

37 38 39
//! Get the maximum allocation size for a machine.
size_t CUDAPinnedMaxAllocSize();

L
liaogang 已提交
40 41 42 43 44
//! Get the minimum chunk size for buddy allocator.
size_t CpuMinChunkSize();

//! Get the maximum chunk size for buddy allocator.
size_t CpuMaxChunkSize();
L
liaogang 已提交
45

46 47 48 49 50 51
//! Get the minimum chunk size for buddy allocator.
size_t CUDAPinnedMinChunkSize();

//! Get the maximum chunk size for buddy allocator.
size_t CUDAPinnedMaxChunkSize();

T
tensor-tang 已提交
52 53 54 55
namespace jit {
typedef enum {
  isa_any,
  sse42,
T
tensor-tang 已提交
56
  avx,
T
tensor-tang 已提交
57
  avx2,
T
tensor-tang 已提交
58
  avx512f,
T
tensor-tang 已提交
59 60 61 62 63 64 65
  avx512_core,
  avx512_core_vnni,
  avx512_mic,
  avx512_mic_4ops,
} cpu_isa_t;  // Instruction set architecture

// May I use some instruction
T
tensor-tang 已提交
66
bool MayIUse(const cpu_isa_t cpu_isa);
T
tensor-tang 已提交
67 68 69

}  // namespace jit

L
liaogang 已提交
70 71
}  // namespace platform
}  // namespace paddle