404 Not Found

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

/*===--------- amxcomplexintrin.h - AMXCOMPLEX intrinsics -*- 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
 *
 *===------------------------------------------------------------------------===
 */

#ifndef __IMMINTRIN_H
#error "Never use <amxcomplexintrin.h> directly; include <immintrin.h> instead."
#endif // __IMMINTRIN_H

#ifndef __AMX_COMPLEXINTRIN_H
#define __AMX_COMPLEXINTRIN_H
#ifdef __x86_64__

#define __DEFAULT_FN_ATTRS_COMPLEX                                             \
  __attribute__((__always_inline__, __nodebug__, __target__("amx-complex")))

/// Perform matrix multiplication of two tiles containing complex elements and
///    accumulate the results into a packed single precision tile. Each dword
///    element in input tiles \a a and \a b is interpreted as a complex number
///    with FP16 real part and FP16 imaginary part.
/// Calculates the imaginary part of the result. For each possible combination
///    of (row of \a a, column of \a b), it performs a set of multiplication
///    and accumulations on all corresponding complex numbers (one from \a a
///    and one from \a b). The imaginary part of the \a a element is multiplied
///    with the real part of the corresponding \a b element, and the real part
///    of the \a a element is multiplied with the imaginary part of the
///    corresponding \a b elements. The two accumulated results are added, and
///    then accumulated into the corresponding row and column of \a dst.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// void _tile_cmmimfp16ps(__tile dst, __tile a, __tile b);
/// \endcode
///
/// \code{.operation}
/// FOR m := 0 TO dst.rows - 1
///	tmp := dst.row[m]
///	FOR k := 0 TO (a.colsb / 4) - 1
///		FOR n := 0 TO (dst.colsb / 4) - 1
///			tmp.fp32[n] += FP32(a.row[m].fp16[2*k+0]) * FP32(b.row[k].fp16[2*n+1])
///			tmp.fp32[n] += FP32(a.row[m].fp16[2*k+1]) * FP32(b.row[k].fp16[2*n+0])
///		ENDFOR
///	ENDFOR
///	write_row_and_zero(dst, m, tmp, dst.colsb)
/// ENDFOR
/// zero_upper_rows(dst, dst.rows)
/// zero_tileconfig_start()
/// \endcode
///
/// This intrinsic corresponds to the \c TCMMIMFP16PS instruction.
///
/// \param dst
///    The destination tile. Max size is 1024 Bytes.
/// \param a
///    The 1st source tile. Max size is 1024 Bytes.
/// \param b
///    The 2nd source tile. Max size is 1024 Bytes.
#define _tile_cmmimfp16ps(dst, a, b) __builtin_ia32_tcmmimfp16ps(dst, a, b)

/// Perform matrix multiplication of two tiles containing complex elements and
///    accumulate the results into a packed single precision tile. Each dword
///    element in input tiles \a a and \a b is interpreted as a complex number
///    with FP16 real part and FP16 imaginary part.
/// Calculates the real part of the result. For each possible combination
///    of (row of \a a, column of \a b), it performs a set of multiplication
///    and accumulations on all corresponding complex numbers (one from \a a
///    and one from \a b). The real part of the \a a element is multiplied
///    with the real part of the corresponding \a b element, and the negated
///    imaginary part of the \a a element is multiplied with the imaginary
///    part of the corresponding \a b elements. The two accumulated results
///    are added, and then accumulated into the corresponding row and column
///    of \a dst.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// void _tile_cmmrlfp16ps(__tile dst, __tile a, __tile b);
/// \endcode
///
/// \code{.operation}
/// FOR m := 0 TO dst.rows - 1
///	tmp := dst.row[m]
///	FOR k := 0 TO (a.colsb / 4) - 1
///		FOR n := 0 TO (dst.colsb / 4) - 1
///			tmp.fp32[n] += FP32(a.row[m].fp16[2*k+0]) * FP32(b.row[k].fp16[2*n+0])
///			tmp.fp32[n] += FP32(-a.row[m].fp16[2*k+1]) * FP32(b.row[k].fp16[2*n+1])
///		ENDFOR
///	ENDFOR
///	write_row_and_zero(dst, m, tmp, dst.colsb)
/// ENDFOR
/// zero_upper_rows(dst, dst.rows)
/// zero_tileconfig_start()
/// \endcode
///
/// This intrinsic corresponds to the \c TCMMIMFP16PS instruction.
///
/// \param dst
///    The destination tile. Max size is 1024 Bytes.
/// \param a
///    The 1st source tile. Max size is 1024 Bytes.
/// \param b
///    The 2nd source tile. Max size is 1024 Bytes.
#define _tile_cmmrlfp16ps(dst, a, b) __builtin_ia32_tcmmrlfp16ps(dst, a, b)

static __inline__ _tile1024i __DEFAULT_FN_ATTRS_COMPLEX
_tile_cmmimfp16ps_internal(unsigned short m, unsigned short n, unsigned short k,
                           _tile1024i dst, _tile1024i src1, _tile1024i src2) {
  return __builtin_ia32_tcmmimfp16ps_internal(m, n, k, dst, src1, src2);
}

static __inline__ _tile1024i __DEFAULT_FN_ATTRS_COMPLEX
_tile_cmmrlfp16ps_internal(unsigned short m, unsigned short n, unsigned short k,
                           _tile1024i dst, _tile1024i src1, _tile1024i src2) {
  return __builtin_ia32_tcmmrlfp16ps_internal(m, n, k, dst, src1, src2);
}

#endif // __x86_64__
#endif // __AMX_COMPLEXINTRIN_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]