source: Daodan/MSYS2/mingw32/lib/gcc/i686-w64-mingw32/11.2.0/include/avx512bf16vlintrin.h

/* Copyright (C) 2019-2021 Free Software Foundation, Inc.

   This file is part of GCC.

   GCC is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 3, or (at your option)
   any later version.

   GCC is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.

   Under Section 7 of GPL version 3, you are granted additional
   permissions described in the GCC Runtime Library Exception, version
   3.1, as published by the Free Software Foundation.

   You should have received a copy of the GNU General Public License and
   a copy of the GCC Runtime Library Exception along with this program;
   see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
   <http://www.gnu.org/licenses/>.  */

#ifndef _IMMINTRIN_H_INCLUDED
#error "Never use <avx512bf16vlintrin.h> directly; include <immintrin.h> instead."
#endif

#ifndef _AVX512BF16VLINTRIN_H_INCLUDED
#define _AVX512BF16VLINTRIN_H_INCLUDED

#if !defined(__AVX512VL__) || !defined(__AVX512BF16__)
#pragma GCC push_options
#pragma GCC target("avx512bf16,avx512vl")
#define __DISABLE_AVX512BF16VL__
#endif /* __AVX512BF16__ */

/* Internal data types for implementing the intrinsics.  */
typedef short __v16bh __attribute__ ((__vector_size__ (32)));
typedef short __v8bh __attribute__ ((__vector_size__ (16)));

/* The Intel API is flexible enough that we must allow aliasing with other
   vector types, and their scalar components.  */
typedef short __m256bh __attribute__ ((__vector_size__ (32), __may_alias__));
typedef short __m128bh __attribute__ ((__vector_size__ (16), __may_alias__));

/* vcvtne2ps2bf16 */

extern __inline __m256bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm256_cvtne2ps_pbh (__m256 __A, __m256 __B)
{
  return (__m256bh)__builtin_ia32_cvtne2ps2bf16_v16hi(__A, __B);
}

extern __inline __m256bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm256_mask_cvtne2ps_pbh (__m256bh __A, __mmask16 __B, __m256 __C, __m256 __D)
{
  return (__m256bh)__builtin_ia32_cvtne2ps2bf16_v16hi_mask(__C, __D, __A, __B);
}

extern __inline __m256bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm256_maskz_cvtne2ps_pbh (__mmask16 __A, __m256 __B, __m256 __C)
{
  return (__m256bh)__builtin_ia32_cvtne2ps2bf16_v16hi_maskz(__B, __C, __A);
}

extern __inline __m128bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtne2ps_pbh (__m128 __A, __m128 __B)
{
  return (__m128bh)__builtin_ia32_cvtne2ps2bf16_v8hi(__A, __B);
}

extern __inline __m128bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_mask_cvtne2ps_pbh (__m128bh __A, __mmask8 __B, __m128 __C, __m128 __D)
{
  return (__m128bh)__builtin_ia32_cvtne2ps2bf16_v8hi_mask(__C, __D, __A, __B);
}

extern __inline __m128bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_maskz_cvtne2ps_pbh (__mmask8 __A, __m128 __B, __m128 __C)
{
  return (__m128bh)__builtin_ia32_cvtne2ps2bf16_v8hi_maskz(__B, __C, __A);
}

/* vcvtneps2bf16 */

extern __inline __m128bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm256_cvtneps_pbh (__m256 __A)
{
  return (__m128bh)__builtin_ia32_cvtneps2bf16_v8sf(__A);
}

extern __inline __m128bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm256_mask_cvtneps_pbh (__m128bh __A, __mmask8 __B, __m256 __C)
{
  return (__m128bh)__builtin_ia32_cvtneps2bf16_v8sf_mask(__C, __A, __B);
}

extern __inline __m128bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm256_maskz_cvtneps_pbh (__mmask8 __A, __m256 __B)
{
  return (__m128bh)__builtin_ia32_cvtneps2bf16_v8sf_maskz(__B, __A);
}

extern __inline __m128bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtneps_pbh (__m128 __A)
{
  return (__m128bh)__builtin_ia32_cvtneps2bf16_v4sf(__A);
}

extern __inline __m128bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_mask_cvtneps_pbh (__m128bh __A, __mmask8 __B, __m128 __C)
{
  return (__m128bh)__builtin_ia32_cvtneps2bf16_v4sf_mask(__C, __A, __B);
}

extern __inline __m128bh
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_maskz_cvtneps_pbh (__mmask8 __A, __m128 __B)
{
  return (__m128bh)__builtin_ia32_cvtneps2bf16_v4sf_maskz(__B, __A);
}

/* vdpbf16ps */

extern __inline __m256
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm256_dpbf16_ps (__m256 __A, __m256bh __B, __m256bh __C)
{
  return (__m256)__builtin_ia32_dpbf16ps_v8sf(__A, __B, __C);
}

extern __inline __m256
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm256_mask_dpbf16_ps (__m256 __A, __mmask8 __B, __m256bh __C, __m256bh __D)
{
  return (__m256)__builtin_ia32_dpbf16ps_v8sf_mask(__A, __C, __D, __B);
}

extern __inline __m256
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm256_maskz_dpbf16_ps (__mmask8 __A, __m256 __B, __m256bh __C, __m256bh __D)
{
  return (__m256)__builtin_ia32_dpbf16ps_v8sf_maskz(__B, __C, __D, __A);
}

extern __inline __m128
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_dpbf16_ps (__m128 __A, __m128bh __B, __m128bh __C)
{
  return (__m128)__builtin_ia32_dpbf16ps_v4sf(__A, __B, __C);
}

extern __inline __m128
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_mask_dpbf16_ps (__m128 __A, __mmask8 __B, __m128bh __C, __m128bh __D)
{
  return (__m128)__builtin_ia32_dpbf16ps_v4sf_mask(__A, __C, __D, __B);
}

extern __inline __m128
__attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_maskz_dpbf16_ps (__mmask8 __A, __m128 __B, __m128bh __C, __m128bh __D)
{
  return (__m128)__builtin_ia32_dpbf16ps_v4sf_maskz(__B, __C, __D, __A);
}

#ifdef __DISABLE_AVX512BF16VL__
#undef __DISABLE_AVX512BF16VL__
#pragma GCC pop_options
#endif /* __DISABLE_AVX512BF16VL__ */

#endif /* _AVX512BF16VLINTRIN_H_INCLUDED */