/external/libaom/libaom/av1/encoder/x86/ |
D | highbd_block_error_intrin_avx2.c | 37 __m256i res = _mm256_mul_epi32(diff1, diff1); in av1_highbd_block_error_avx2() 38 __m256i res1 = _mm256_mul_epi32(diff1h, diff1h); in av1_highbd_block_error_avx2() 39 __m256i res2 = _mm256_mul_epi32(diff2, diff2); in av1_highbd_block_error_avx2() 40 __m256i res3 = _mm256_mul_epi32(diff2h, diff2h); in av1_highbd_block_error_avx2() 45 res = _mm256_mul_epi32(mm256_coeff, mm256_coeff); in av1_highbd_block_error_avx2() 46 res1 = _mm256_mul_epi32(mm256_coeffh, mm256_coeffh); in av1_highbd_block_error_avx2() 47 res2 = _mm256_mul_epi32(mm256_coeff2, mm256_coeff2); in av1_highbd_block_error_avx2() 48 res3 = _mm256_mul_epi32(mm256_coeffh2, mm256_coeffh2); in av1_highbd_block_error_avx2()
|
D | pickrst_avx2.c | 660 const __m256i h00_even = _mm256_mul_epi32(f1, f1); in calc_proj_params_r0_r1_avx2() 661 const __m256i h00_odd = _mm256_mul_epi32(_mm256_srli_epi64(f1, 32), in calc_proj_params_r0_r1_avx2() 666 const __m256i h01_even = _mm256_mul_epi32(f1, f2); in calc_proj_params_r0_r1_avx2() 667 const __m256i h01_odd = _mm256_mul_epi32(_mm256_srli_epi64(f1, 32), in calc_proj_params_r0_r1_avx2() 672 const __m256i h11_even = _mm256_mul_epi32(f2, f2); in calc_proj_params_r0_r1_avx2() 673 const __m256i h11_odd = _mm256_mul_epi32(_mm256_srli_epi64(f2, 32), in calc_proj_params_r0_r1_avx2() 678 const __m256i c0_even = _mm256_mul_epi32(f1, s); in calc_proj_params_r0_r1_avx2() 680 _mm256_mul_epi32(_mm256_srli_epi64(f1, 32), _mm256_srli_epi64(s, 32)); in calc_proj_params_r0_r1_avx2() 684 const __m256i c1_even = _mm256_mul_epi32(f2, s); in calc_proj_params_r0_r1_avx2() 686 _mm256_mul_epi32(_mm256_srli_epi64(f2, 32), _mm256_srli_epi64(s, 32)); in calc_proj_params_r0_r1_avx2() [all …]
|
D | av1_highbd_quantize_avx2.c | 55 __m256i q_lo = _mm256_mul_epi32(q, qp[1]); in quantize() 58 q_hi = _mm256_mul_epi32(q_hi, qp_hi); in quantize()
|
/external/flac/src/libFLAC/ |
D | lpc_intrin_avx2.c | 794 … summ = _mm256_mul_epi32(q11, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-12)))); in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() 795 …mull = _mm256_mul_epi32(q10, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-11))));… in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() 796 …mull = _mm256_mul_epi32(q9, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-10))));… in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() 797 …mull = _mm256_mul_epi32(q8, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-9 ))));… in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() 798 …mull = _mm256_mul_epi32(q7, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-8 ))));… in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() 799 …mull = _mm256_mul_epi32(q6, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-7 ))));… in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() 800 …mull = _mm256_mul_epi32(q5, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-6 ))));… in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() 801 …mull = _mm256_mul_epi32(q4, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-5 ))));… in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() 802 …mull = _mm256_mul_epi32(q3, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-4 ))));… in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() 803 …mull = _mm256_mul_epi32(q2, _mm256_cvtepu32_epi64(_mm_loadu_si128((const __m128i*)(data+i-3 ))));… in FLAC__lpc_compute_residual_from_qlp_coefficients_wide_intrin_avx2() [all …]
|
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 3x8c8-minmax-avx2.c | 160 …const __m256i vprod0x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x01234567, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2() 161 …const __m256i vprod1x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x01234567, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2() 162 …const __m256i vprod2x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc2x01234567, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2() 164 …const __m256i vprod0x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x11335577, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2() 165 …const __m256i vprod1x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x11335577, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2() 166 …const __m256i vprod2x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc2x11335577, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2()
|
D | 2x8c8-minmax-avx2.c | 136 …const __m256i vprod0x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x01234567, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_2x8c8__avx2() 137 …const __m256i vprod1x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x01234567, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_2x8c8__avx2() 139 …const __m256i vprod0x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x11335577, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_2x8c8__avx2() 140 …const __m256i vprod1x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x11335577, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_2x8c8__avx2()
|
D | 1x8c8-minmax-avx2.c | 112 …const __m256i vprod0x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x01234567, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_1x8c8__avx2() 114 …const __m256i vprod0x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x11335577, vmultiplier), vround… in xnn_qs8_igemm_minmax_ukernel_1x8c8__avx2()
|
/external/XNNPACK/src/qs8-gemm/gen/ |
D | 3x8c8-minmax-avx2.c | 143 …const __m256i vprod0x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x01234567, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2() 144 …const __m256i vprod1x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x01234567, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2() 145 …const __m256i vprod2x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc2x01234567, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2() 147 …const __m256i vprod0x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x11335577, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2() 148 …const __m256i vprod1x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x11335577, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2() 149 …const __m256i vprod2x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc2x11335577, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2()
|
D | 3x8c8-xw-minmax-avx2.c | 139 …const __m256i vprod0x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x01234567, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2() 140 …const __m256i vprod1x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x01234567, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2() 141 …const __m256i vprod2x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc2x01234567, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2() 143 …const __m256i vprod0x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x11335577, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2() 144 …const __m256i vprod1x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x11335577, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2() 145 …const __m256i vprod2x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc2x11335577, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2()
|
D | 2x8c8-minmax-avx2.c | 121 …const __m256i vprod0x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x01234567, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_2x8c8__avx2() 122 …const __m256i vprod1x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x01234567, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_2x8c8__avx2() 124 …const __m256i vprod0x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x11335577, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_2x8c8__avx2() 125 …const __m256i vprod1x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x11335577, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_2x8c8__avx2()
|
D | 2x8c8-xw-minmax-avx2.c | 117 …const __m256i vprod0x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x01234567, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_2x8c8__avx2() 118 …const __m256i vprod1x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x01234567, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_2x8c8__avx2() 120 …const __m256i vprod0x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x11335577, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_2x8c8__avx2() 121 …const __m256i vprod1x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1x11335577, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_2x8c8__avx2()
|
D | 1x8c8-minmax-avx2.c | 99 …const __m256i vprod0x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x01234567, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_1x8c8__avx2() 101 …const __m256i vprod0x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x11335577, vmultiplier), vround… in xnn_qs8_gemm_minmax_ukernel_1x8c8__avx2()
|
D | 1x8c8-xw-minmax-avx2.c | 95 …const __m256i vprod0x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x01234567, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_1x8c8__avx2() 97 …const __m256i vprod0x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc0x11335577, vmultiplier), vround… in xnn_qs8_gemm_xw_minmax_ukernel_1x8c8__avx2()
|
/external/XNNPACK/src/qs8-dwconv/gen/ |
D | up32x9-minmax-avx2-mul16.c | 234 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() 235 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() 236 …const __m256i vprod8ACE = _mm256_add_epi64(_mm256_mul_epi32(vacc89ABCDEF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() 237 … const __m256i vprod9BDF = _mm256_add_epi64(_mm256_mul_epi32(vacc9BDF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() 238 …const __m256i vprodGIKM = _mm256_add_epi64(_mm256_mul_epi32(vaccGHIJKLMN, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() 239 … const __m256i vprodHJLN = _mm256_add_epi64(_mm256_mul_epi32(vaccHJLN, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() 240 …const __m256i vprodOQSU = _mm256_add_epi64(_mm256_mul_epi32(vaccOPQRSTUV, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() 241 … const __m256i vprodPRTV = _mm256_add_epi64(_mm256_mul_epi32(vaccPRTV, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() 391 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() 392 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul16() [all …]
|
D | up16x9-minmax-avx2-mul16.c | 176 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul16() 177 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul16() 178 …const __m256i vprod8ACE = _mm256_add_epi64(_mm256_mul_epi32(vacc89ABCDEF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul16() 179 … const __m256i vprod9BDF = _mm256_add_epi64(_mm256_mul_epi32(vacc9BDF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul16() 299 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul16() 300 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul16() 301 …const __m256i vprod8ACE = _mm256_add_epi64(_mm256_mul_epi32(vacc89ABCDEF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul16() 302 … const __m256i vprod9BDF = _mm256_add_epi64(_mm256_mul_epi32(vacc9BDF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul16()
|
D | up32x9-minmax-avx2-mul32.c | 234 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 235 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 236 …const __m256i vprod8ACE = _mm256_add_epi64(_mm256_mul_epi32(vacc89ABCDEF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 237 … const __m256i vprod9BDF = _mm256_add_epi64(_mm256_mul_epi32(vacc9BDF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 238 …const __m256i vprodGIKM = _mm256_add_epi64(_mm256_mul_epi32(vaccGHIJKLMN, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 239 … const __m256i vprodHJLN = _mm256_add_epi64(_mm256_mul_epi32(vaccHJLN, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 240 …const __m256i vprodOQSU = _mm256_add_epi64(_mm256_mul_epi32(vaccOPQRSTUV, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 241 … const __m256i vprodPRTV = _mm256_add_epi64(_mm256_mul_epi32(vaccPRTV, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 362 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 363 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32()
|
D | up24x9-minmax-avx2-mul32.c | 205 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 206 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 207 …const __m256i vprod8ACE = _mm256_add_epi64(_mm256_mul_epi32(vacc89ABCDEF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 208 … const __m256i vprod9BDF = _mm256_add_epi64(_mm256_mul_epi32(vacc9BDF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 209 …const __m256i vprodGIKM = _mm256_add_epi64(_mm256_mul_epi32(vaccGHIJKLMN, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 210 … const __m256i vprodHJLN = _mm256_add_epi64(_mm256_mul_epi32(vaccHJLN, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 324 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 325 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32()
|
D | up16x9-minmax-avx2-mul32.c | 176 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 177 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 178 …const __m256i vprod8ACE = _mm256_add_epi64(_mm256_mul_epi32(vacc89ABCDEF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 179 … const __m256i vprod9BDF = _mm256_add_epi64(_mm256_mul_epi32(vacc9BDF, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 282 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 283 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32()
|
D | up8x9-minmax-avx2-mul32.c | 147 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 148 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 232 …const __m256i vprod0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc01234567, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 233 … const __m256i vprod1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc1357, vmultiplier), vrounding); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32()
|
/external/XNNPACK/src/qs8-dwconv/ |
D | unipass-avx2-mul16.c.in | 73 …const __m256i vprod${ABC[C:C+8:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[C:C+8]}, vmultipl… 74 …const __m256i vprod${ABC[C+1:C+8:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[C+1:C+8:2]}, vm… 146 …const __m256i vprod${ABC[0:8:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[0:8]}, vmultiplier)… 147 …const __m256i vprod${ABC[1:8:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[1:8:2]}, vmultiplie… 148 …const __m256i vprod${ABC[8:16:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[8:16]}, vmultiplie… 149 …const __m256i vprod${ABC[9:16:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[9:16:2]}, vmultipl…
|
D | unipass-avx2-mul32.c.in | 70 …const __m256i vprod${ABC[C:C+8:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[C:C+8]}, vmultipl… 71 …const __m256i vprod${ABC[C+1:C+8:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[C+1:C+8:2]}, vm… 164 …const __m256i vprod${ABC[0:8:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[0:8]}, vmultiplier)… 165 …const __m256i vprod${ABC[1:8:2]} = _mm256_add_epi64(_mm256_mul_epi32(vacc${ABC[1:8:2]}, vmultiplie…
|
/external/libaom/libaom/aom_dsp/x86/ |
D | highbd_quantize_intrin_avx2.c | 52 __m256i prod_lo = _mm256_mul_epi32(*x, *y); in mm256_mul_shift_epi32() 55 prod_hi = _mm256_mul_epi32(prod_hi, mult_hi); in mm256_mul_shift_epi32()
|
/external/gemmlowp/fixedpoint/ |
D | fixedpoint_avx.h | 334 a0b0_a2b2 = _mm256_mul_epi32(a0_a2, b0_b2); 335 a1b1_a3b3 = _mm256_mul_epi32(a1_a3, b1_b3);
|
/external/XNNPACK/src/qs8-igemm/ |
D | MRx8c8-avx2.c.in | 121 …const __m256i vprod${M}x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc${M}x01234567, vmultiplier), … 124 …const __m256i vprod${M}x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc${M}x11335577, vmultiplier), …
|
/external/XNNPACK/src/qs8-gemm/ |
D | MRx8c8-avx2.c.in | 122 …const __m256i vprod${M}x0246 = _mm256_add_epi64(_mm256_mul_epi32(vacc${M}x01234567, vmultiplier), … 125 …const __m256i vprod${M}x1357 = _mm256_add_epi64(_mm256_mul_epi32(vacc${M}x11335577, vmultiplier), …
|