404 Not Found

PATH: lib64 / llvm17 / lib / clang / 17 / include

/*===---- ammintrin.h - SSE4a intrinsics -----------------------------------===
 *
 * 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
 *
 *===-----------------------------------------------------------------------===
 */

#ifndef __AMMINTRIN_H
#define __AMMINTRIN_H

#if !defined(__i386__) && !defined(__x86_64__)
#error "This header is only meant to be used on x86 and x64 architecture"
#endif

#include <pmmintrin.h>

/* Define the default attributes for the functions in this file. */
#define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("sse4a"), __min_vector_width__(128)))

/// Extracts the specified bits from the lower 64 bits of the 128-bit
///    integer vector operand at the index \a idx and of the length \a len.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// __m128i _mm_extracti_si64(__m128i x, const int len, const int idx);
/// \endcode
///
/// This intrinsic corresponds to the <c> EXTRQ </c> instruction.
///
/// \param x
///    The value from which bits are extracted.
/// \param len
///    Bits [5:0] specify the length; the other bits are ignored. If bits [5:0]
///    are zero, the length is interpreted as 64.
/// \param idx
///    Bits [5:0] specify the index of the least significant bit; the other
///    bits are ignored. If the sum of the index and length is greater than 64,
///    the result is undefined. If the length and index are both zero, bits
///    [63:0] of parameter \a x are extracted. If the length is zero but the
///    index is non-zero, the result is undefined.
/// \returns A 128-bit integer vector whose lower 64 bits contain the bits
///    extracted from the source operand.
#define _mm_extracti_si64(x, len, idx) \
  ((__m128i)__builtin_ia32_extrqi((__v2di)(__m128i)(x), \
                                  (char)(len), (char)(idx)))

/// Extracts the specified bits from the lower 64 bits of the 128-bit
///    integer vector operand at the index and of the length specified by
///    \a __y.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> EXTRQ </c> instruction.
///
/// \param __x
///    The value from which bits are extracted.
/// \param __y
///    Specifies the index of the least significant bit at [13:8] and the
///    length at [5:0]; all other bits are ignored. If bits [5:0] are zero, the
///    length is interpreted as 64. If the sum of the index and length is
///    greater than 64, the result is undefined. If the length and index are
///    both zero, bits [63:0] of parameter \a __x are extracted. If the length
///    is zero but the index is non-zero, the result is undefined.
/// \returns A 128-bit vector whose lower 64 bits contain the bits extracted
///    from the source operand.
static __inline__ __m128i __DEFAULT_FN_ATTRS
_mm_extract_si64(__m128i __x, __m128i __y)
{
  return (__m128i)__builtin_ia32_extrq((__v2di)__x, (__v16qi)__y);
}

/// Inserts bits of a specified length from the source integer vector
///    \a y into the lower 64 bits of the destination integer vector \a x at
///    the index \a idx and of the length \a len.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// __m128i _mm_inserti_si64(__m128i x, __m128i y, const int len,
/// const int idx);
/// \endcode
///
/// This intrinsic corresponds to the <c> INSERTQ </c> instruction.
///
/// \param x
///    The destination operand where bits will be inserted. The inserted bits
///    are defined by the length \a len and by the index \a idx specifying the
///    least significant bit.
/// \param y
///    The source operand containing the bits to be extracted. The extracted
///    bits are the least significant bits of operand \a y of length \a len.
/// \param len
///    Bits [5:0] specify the length; the other bits are ignored. If bits [5:0]
///    are zero, the length is interpreted as 64.
/// \param idx
///    Bits [5:0] specify the index of the least significant bit; the other
///    bits are ignored. If the sum of the index and length is greater than 64,
///    the result is undefined. If the length and index are both zero, bits
///    [63:0] of parameter \a y are inserted into parameter \a x. If the length
///    is zero but the index is non-zero, the result is undefined.
/// \returns A 128-bit integer vector containing the original lower 64-bits of
///    destination operand \a x with the specified bitfields replaced by the
///    lower bits of source operand \a y. The upper 64 bits of the return value
///    are undefined.
#define _mm_inserti_si64(x, y, len, idx) \
  ((__m128i)__builtin_ia32_insertqi((__v2di)(__m128i)(x), \
                                    (__v2di)(__m128i)(y), \
                                    (char)(len), (char)(idx)))

/// Inserts bits of a specified length from the source integer vector
///    \a __y into the lower 64 bits of the destination integer vector \a __x
///    at the index and of the length specified by \a __y.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> INSERTQ </c> instruction.
///
/// \param __x
///    The destination operand where bits will be inserted. The inserted bits
///    are defined by the length and by the index of the least significant bit
///    specified by operand \a __y.
/// \param __y
///    The source operand containing the bits to be extracted. The extracted
///    bits are the least significant bits of operand \a __y with length
///    specified by bits [69:64]. These are inserted into the destination at the
///    index specified by bits [77:72]; all other bits are ignored. If bits
///    [69:64] are zero, the length is interpreted as 64. If the sum of the
///    index and length is greater than 64, the result is undefined. If the
///    length and index are both zero, bits [63:0] of parameter \a __y are
///    inserted into parameter \a __x. If the length is zero but the index is
///    non-zero, the result is undefined.
/// \returns A 128-bit integer vector containing the original lower 64-bits of
///    destination operand \a __x with the specified bitfields replaced by the
///    lower bits of source operand \a __y. The upper 64 bits of the return
///    value are undefined.
static __inline__ __m128i __DEFAULT_FN_ATTRS
_mm_insert_si64(__m128i __x, __m128i __y)
{
  return (__m128i)__builtin_ia32_insertq((__v2di)__x, (__v2di)__y);
}

/// Stores a 64-bit double-precision value in a 64-bit memory location.
///    To minimize caching, the data is flagged as non-temporal (unlikely to be
///    used again soon).
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> MOVNTSD </c> instruction.
///
/// \param __p
///    The 64-bit memory location used to store the register value.
/// \param __a
///    The 64-bit double-precision floating-point register value to be stored.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_stream_sd(double *__p, __m128d __a)
{
  __builtin_ia32_movntsd(__p, (__v2df)__a);
}

/// Stores a 32-bit single-precision floating-point value in a 32-bit
///    memory location. To minimize caching, the data is flagged as
///    non-temporal (unlikely to be used again soon).
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> MOVNTSS </c> instruction.
///
/// \param __p
///    The 32-bit memory location used to store the register value.
/// \param __a
///    The 32-bit single-precision floating-point register value to be stored.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_stream_ss(float *__p, __m128 __a)
{
  __builtin_ia32_movntss(__p, (__v4sf)__a);
}

#undef __DEFAULT_FN_ATTRS

#endif /* __AMMINTRIN_H */

[+] ..
[-] riscv_ntlh.h [edit]
[-] avx512vlbitalgintrin.h [edit]
[-] avxifmaintrin.h [edit]
[-] mm_malloc.h [edit]
[-] avx512vlbf16intrin.h [edit]
[-] mwaitxintrin.h [edit]
[-] adxintrin.h [edit]
[-] arm_cmse.h [edit]
[-] __clang_cuda_intrinsics.h [edit]
[-] module.modulemap [edit]
[-] crc32intrin.h [edit]
[+] cuda_wrappers
[-] msa.h [edit]
[-] __clang_hip_stdlib.h [edit]
[-] arm_acle.h [edit]
[-] arm_neon_sve_bridge.h [edit]
[-] xsaveintrin.h [edit]
[-] velintrin_gen.h [edit]
[-] stdatomic.h [edit]
[-] hvx_hexagon_protos.h [edit]
[-] uintrintrin.h [edit]
[-] fxsrintrin.h [edit]
[-] sha512intrin.h [edit]
[-] __clang_cuda_texture_intrinsics.h [edit]
[-] iso646.h [edit]
[-] unwind.h [edit]
[-] avx512vlbwintrin.h [edit]
[-] avxvnniint8intrin.h [edit]
[-] avx512erintrin.h [edit]
[-] avxvnniintrin.h [edit]
[-] ia32intrin.h [edit]
[-] rdseedintrin.h [edit]
[-] prfchiintrin.h [edit]
[-] amxcomplexintrin.h [edit]
[-] clflushoptintrin.h [edit]
[-] htmxlintrin.h [edit]
[-] avx512fintrin.h [edit]
[-] gfniintrin.h [edit]
[-] arm_cde.h [edit]
[+] ppc_wrappers
[-] amxfp16intrin.h [edit]
[-] velintrin_approx.h [edit]
[-] avx512pfintrin.h [edit]
[-] stdarg.h [edit]
[-] cmpccxaddintrin.h [edit]
[-] avx512vlintrin.h [edit]
[-] __clang_hip_math.h [edit]
[-] vecintrin.h [edit]
[-] xtestintrin.h [edit]
[-] __wmmintrin_aes.h [edit]
[-] arm_neon.h [edit]
[-] immintrin.h [edit]
[-] ammintrin.h [edit]
[-] waitpkgintrin.h [edit]
[-] vpclmulqdqintrin.h [edit]
[-] fmaintrin.h [edit]
[-] tsxldtrkintrin.h [edit]
[-] prfchwintrin.h [edit]
[-] avx512bitalgintrin.h [edit]
[-] bmiintrin.h [edit]
[-] __wmmintrin_pclmul.h [edit]
[-] htmintrin.h [edit]
[-] mm3dnow.h [edit]
[-] __clang_cuda_builtin_vars.h [edit]
[-] __clang_hip_runtime_wrapper.h [edit]
[-] stdbool.h [edit]
[-] altivec.h [edit]
[-] wbnoinvdintrin.h [edit]
[-] keylockerintrin.h [edit]
[-] tgmath.h [edit]
[-] hexagon_circ_brev_intrinsics.h [edit]
[-] x86intrin.h [edit]
[-] pkuintrin.h [edit]
[-] avx512vbmivlintrin.h [edit]
[-] avxneconvertintrin.h [edit]
[-] __clang_hip_cmath.h [edit]
[-] sgxintrin.h [edit]
[-] f16cintrin.h [edit]
[-] opencl-c-base.h [edit]
[-] cpuid.h [edit]
[-] raointintrin.h [edit]
[-] builtins.h [edit]
[-] emmintrin.h [edit]
[-] smmintrin.h [edit]
[-] vaesintrin.h [edit]
[-] larchintrin.h [edit]
[-] avx512ifmaintrin.h [edit]
[-] intrin.h [edit]
[-] avx512vlvp2intersectintrin.h [edit]
[-] fma4intrin.h [edit]
[-] pmmintrin.h [edit]
[-] __clang_hip_libdevice_declares.h [edit]
[-] limits.h [edit]
[-] clwbintrin.h [edit]
[-] rtmintrin.h [edit]
[-] mmintrin.h [edit]
[-] stddef.h [edit]
[-] invpcidintrin.h [edit]
[-] avx512vp2intersectintrin.h [edit]
[-] cet.h [edit]
[-] xopintrin.h [edit]
[-] avx512vlvnniintrin.h [edit]
[-] avx512vlfp16intrin.h [edit]
[-] stdint.h [edit]
[-] arm64intr.h [edit]
[-] sm4intrin.h [edit]
[-] avx512vnniintrin.h [edit]
[-] avx2intrin.h [edit]
[-] movdirintrin.h [edit]
[-] tbmintrin.h [edit]
[-] arm_mve.h [edit]
[-] avx512ifmavlintrin.h [edit]
[-] amxintrin.h [edit]
[-] opencl-c.h [edit]
[-] stdalign.h [edit]
[-] __clang_cuda_device_functions.h [edit]
[-] pconfigintrin.h [edit]
[-] avx512fp16intrin.h [edit]
[-] inttypes.h [edit]
[-] arm_bf16.h [edit]
[-] __clang_cuda_math_forward_declares.h [edit]
[-] vadefs.h [edit]
[-] shaintrin.h [edit]
[-] hexagon_protos.h [edit]
[-] ptwriteintrin.h [edit]
[-] xsaveoptintrin.h [edit]
[-] enqcmdintrin.h [edit]
[-] x86gprintrin.h [edit]
[-] tmmintrin.h [edit]
[-] stdnoreturn.h [edit]
[-] avx512bf16intrin.h [edit]
[-] varargs.h [edit]
[-] s390intrin.h [edit]
[-] avx512vbmiintrin.h [edit]
[+] openmp_wrappers
[-] wmmintrin.h [edit]
[-] __clang_cuda_cmath.h [edit]
[-] clzerointrin.h [edit]
[-] xsavesintrin.h [edit]
[-] __clang_cuda_libdevice_declares.h [edit]
[-] nmmintrin.h [edit]
[-] wasm_simd128.h [edit]
[-] xsavecintrin.h [edit]
[-] avx512dqintrin.h [edit]
[-] lwpintrin.h [edit]
[-] serializeintrin.h [edit]
[-] arm_sme_draft_spec_subject_to_change.h [edit]
[-] avxintrin.h [edit]
[-] __stddef_max_align_t.h [edit]
[-] sm3intrin.h [edit]
[-] velintrin.h [edit]
[-] __clang_cuda_complex_builtins.h [edit]
[-] armintr.h [edit]
[-] avx512cdintrin.h [edit]
[-] float.h [edit]
[-] avx512vbmi2intrin.h [edit]
[-] lzcntintrin.h [edit]
[-] sifive_vector.h [edit]
[-] rdpruintrin.h [edit]
[-] arm_sve.h [edit]
[-] avx512vpopcntdqintrin.h [edit]
[-] xmmintrin.h [edit]
[-] hresetintrin.h [edit]
[-] bmi2intrin.h [edit]
[-] hexagon_types.h [edit]
[-] avx512bwintrin.h [edit]
[-] cetintrin.h [edit]
[-] __clang_cuda_math.h [edit]
[-] avx512vlvbmi2intrin.h [edit]
[-] arm_fp16.h [edit]
[-] avx512vpopcntdqvlintrin.h [edit]
[-] avxvnniint16intrin.h [edit]
[+] llvm_libc_wrappers
[-] __clang_cuda_runtime_wrapper.h [edit]
[-] cldemoteintrin.h [edit]
[-] avx512vldqintrin.h [edit]
[-] avx512vlcdintrin.h [edit]
[-] popcntintrin.h [edit]