/usr/local/lib64/python3.6/site-packages/torch/include/c10/util
Edit: /usr/local/lib64/python3.6/site-packages/torch/include/c10/util/BFloat16.h (2372B)
#pragma once
// Defines the bloat16 type (brain floating-point). This representation uses
// 1 bit for the sign, 8 bits for the exponent and 7 bits for the mantissa.
#include
#include
#include
#if defined(CUDA_VERSION) && CUDA_VERSION >= 11000
#include
#endif
namespace c10 {
namespace detail {
inline C10_HOST_DEVICE float f32_from_bits(uint16_t src) {
float res = 0;
uint32_t tmp = src;
tmp <<= 16;
#ifdef __HIP_PLATFORM_HCC__
float* tempRes;
// We should be using memcpy in order to respect the strict aliasing rule
// but it fails in the HIP environment.
tempRes = reinterpret_cast(&tmp);
res = *tempRes;
#else
std::memcpy(&res, &tmp, sizeof(tmp));
#endif
return res;
}
inline C10_HOST_DEVICE uint16_t bits_from_f32(float src) {
uint32_t res = 0;
#ifdef __HIP_PLATFORM_HCC__
// We should be using memcpy in order to respect the strict aliasing rule
// but it fails in the HIP environment.
uint32_t* tempRes = reinterpret_cast(&src);
res = *tempRes;
#else
std::memcpy(&res, &src, sizeof(res));
#endif
return res >> 16;
}
inline C10_HOST_DEVICE uint16_t round_to_nearest_even(float src) {
#if defined(__HIP_PLATFORM_HCC__)
if (src != src) {
#elif defined(_MSC_VER)
if (isnan(src)) {
#else
if (std::isnan(src)) {
#endif
return UINT16_C(0x7FC0);
} else {
union {
uint32_t U32;
float F32;
};
F32 = src;
uint32_t rounding_bias = ((U32 >> 16) & 1) + UINT32_C(0x7FFF);
return static_cast((U32 + rounding_bias) >> 16);
}
}
} // namespace detail
struct alignas(2) BFloat16 {
uint16_t x;
// HIP wants __host__ __device__ tag, CUDA does not
#ifdef __HIP_PLATFORM_HCC__
C10_HOST_DEVICE BFloat16() = default;
#else
BFloat16() = default;
#endif
struct from_bits_t {};
static constexpr C10_HOST_DEVICE from_bits_t from_bits() {
return from_bits_t();
}
constexpr C10_HOST_DEVICE BFloat16(unsigned short bits, from_bits_t)
: x(bits){};
inline C10_HOST_DEVICE BFloat16(float value);
inline C10_HOST_DEVICE operator float() const;
#if defined(CUDA_VERSION) && CUDA_VERSION >= 11000
inline C10_HOST_DEVICE BFloat16(const __nv_bfloat16& value);
explicit inline C10_HOST_DEVICE operator __nv_bfloat16() const;
#endif
};
} // namespace c10
#include