/external/libaom/libaom/av1/encoder/x86/ |
D | av1_fwd_txfm_avx2.h | 25 const __m256i in0_w0 = _mm256_mullo_epi32(_in0, ww0); in btf_32_avx2_type0() 26 const __m256i in1_w1 = _mm256_mullo_epi32(_in1, ww1); in btf_32_avx2_type0() 30 const __m256i in0_w1 = _mm256_mullo_epi32(_in0, ww1); in btf_32_avx2_type0() 31 const __m256i in1_w0 = _mm256_mullo_epi32(_in1, ww0); in btf_32_avx2_type0() 44 const __m256i in0_w0 = _mm256_mullo_epi32(_in0, ww0); in btf_32_avx2_type1() 45 const __m256i in1_w1 = _mm256_mullo_epi32(_in1, ww1); in btf_32_avx2_type1() 49 const __m256i in0_w1 = _mm256_mullo_epi32(_in0, ww1); in btf_32_avx2_type1() 50 const __m256i in1_w0 = _mm256_mullo_epi32(_in1, ww0); in btf_32_avx2_type1() 64 const __m256i in0_w0 = _mm256_mullo_epi32(_in0, ww0); in btf_32_avx2_type0_new() 65 const __m256i in1_w1 = _mm256_mullo_epi32(_in1, ww1); in btf_32_avx2_type0_new() [all …]
|
D | highbd_fwd_txfm_avx2.c | 217 x = _mm256_mullo_epi32(*w0, *n0); in av1_half_btf_avx2() 218 y = _mm256_mullo_epi32(*w1, *n1); in av1_half_btf_avx2() 228 const __m256i in0_w0 = _mm256_mullo_epi32(in0, ww0); \ 229 const __m256i in1_w1 = _mm256_mullo_epi32(in1, ww1); \ 232 const __m256i in0_w1 = _mm256_mullo_epi32(in0, ww1); \ 233 const __m256i in1_w0 = _mm256_mullo_epi32(in1, ww0); \ 240 const __m256i in0_w0 = _mm256_mullo_epi32(in0, ww0); \ 241 const __m256i in1_w1 = _mm256_mullo_epi32(in1, ww1); \ 245 const __m256i in0_w1 = _mm256_mullo_epi32(in0, ww1); \ 246 const __m256i in1_w0 = _mm256_mullo_epi32(in1, ww0); \ [all …]
|
D | pickrst_avx2.c | 910 const __m256i v0l = _mm256_mullo_epi32(flt0l_subu, xq0); in av1_highbd_pixel_proj_error_avx2() 911 const __m256i v0h = _mm256_mullo_epi32(flt0h_subu, xq0); in av1_highbd_pixel_proj_error_avx2() 912 const __m256i v1l = _mm256_mullo_epi32(flt1l_subu, xq1); in av1_highbd_pixel_proj_error_avx2() 913 const __m256i v1h = _mm256_mullo_epi32(flt1h_subu, xq1); in av1_highbd_pixel_proj_error_avx2() 988 const __m256i flth_xq = _mm256_mullo_epi32(flth, xq_active); in av1_highbd_pixel_proj_error_avx2() 989 const __m256i fltl_xq = _mm256_mullo_epi32(fltl, xq_active); in av1_highbd_pixel_proj_error_avx2() 990 const __m256i d0h_xq = _mm256_mullo_epi32(d0h, xq_inactive); in av1_highbd_pixel_proj_error_avx2() 991 const __m256i d0l_xq = _mm256_mullo_epi32(d0l, xq_inactive); in av1_highbd_pixel_proj_error_avx2()
|
D | av1_highbd_quantize_avx2.c | 67 __m256i dq = _mm256_mullo_epi32(q, qp[2]); in quantize()
|
/external/XNNPACK/src/qs8-dwconv/gen/ |
D | up32x9-minmax-avx2-mul32.c | 99 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 100 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 101 vaccGHIJKLMN = _mm256_add_epi32(vaccGHIJKLMN, _mm256_mullo_epi32(vi0xGHIJKLMN, vk0xGHIJKLMN)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 102 vaccOPQRSTUV = _mm256_add_epi32(vaccOPQRSTUV, _mm256_mullo_epi32(vi0xOPQRSTUV, vk0xOPQRSTUV)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 114 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 115 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 116 vaccGHIJKLMN = _mm256_add_epi32(vaccGHIJKLMN, _mm256_mullo_epi32(vi1xGHIJKLMN, vk1xGHIJKLMN)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 117 vaccOPQRSTUV = _mm256_add_epi32(vaccOPQRSTUV, _mm256_mullo_epi32(vi1xOPQRSTUV, vk1xOPQRSTUV)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 129 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() 130 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up32x9__avx2_mul32() [all …]
|
D | up24x9-minmax-avx2-mul32.c | 96 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 97 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 98 vaccGHIJKLMN = _mm256_add_epi32(vaccGHIJKLMN, _mm256_mullo_epi32(vi0xGHIJKLMN, vk0xGHIJKLMN)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 108 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 109 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 110 vaccGHIJKLMN = _mm256_add_epi32(vaccGHIJKLMN, _mm256_mullo_epi32(vi1xGHIJKLMN, vk1xGHIJKLMN)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 120 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 121 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 122 vaccGHIJKLMN = _mm256_add_epi32(vaccGHIJKLMN, _mm256_mullo_epi32(vi2xGHIJKLMN, vk2xGHIJKLMN)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() 132 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__avx2_mul32() [all …]
|
D | up16x9-minmax-avx2-mul32.c | 93 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 94 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 102 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 103 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 111 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 112 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 120 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 121 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi3x89ABCDEF, vk3x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 129 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() 130 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi4x89ABCDEF, vk4x89ABCDEF)); in xnn_qs8_dwconv_minmax_ukernel_up16x9__avx2_mul32() [all …]
|
D | up8x9-minmax-avx2-mul32.c | 90 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 96 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 102 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 108 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 114 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 120 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 126 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 132 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 138 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() 184 … vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567)); in xnn_qs8_dwconv_minmax_ukernel_up8x9__avx2_mul32() [all …]
|
/external/flac/src/libFLAC/ |
D | lpc_intrin_avx2.c | 435 summ = _mm256_mullo_epi32(q11, _mm256_loadu_si256((const __m256i*)(data+i-12))); in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() 436 …mull = _mm256_mullo_epi32(q10, _mm256_loadu_si256((const __m256i*)(data+i-11))); summ = _mm256_add… in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() 437 …mull = _mm256_mullo_epi32(q9, _mm256_loadu_si256((const __m256i*)(data+i-10))); summ = _mm256_add… in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() 438 …mull = _mm256_mullo_epi32(q8, _mm256_loadu_si256((const __m256i*)(data+i-9))); summ = _mm256_add… in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() 439 …mull = _mm256_mullo_epi32(q7, _mm256_loadu_si256((const __m256i*)(data+i-8))); summ = _mm256_add… in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() 440 …mull = _mm256_mullo_epi32(q6, _mm256_loadu_si256((const __m256i*)(data+i-7))); summ = _mm256_add… in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() 441 …mull = _mm256_mullo_epi32(q5, _mm256_loadu_si256((const __m256i*)(data+i-6))); summ = _mm256_add… in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() 442 …mull = _mm256_mullo_epi32(q4, _mm256_loadu_si256((const __m256i*)(data+i-5))); summ = _mm256_add… in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() 443 …mull = _mm256_mullo_epi32(q3, _mm256_loadu_si256((const __m256i*)(data+i-4))); summ = _mm256_add… in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() 444 …mull = _mm256_mullo_epi32(q2, _mm256_loadu_si256((const __m256i*)(data+i-3))); summ = _mm256_add… in FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_avx2() [all …]
|
/external/XNNPACK/src/qs8-vadd/gen/ |
D | minmax-avx2-mul32-ld64-x32.c | 47 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 48 …__m256i vacc89ABCDEF = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx89ABCDEF, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 49 …__m256i vaccGHIJKLMN = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vxGHIJKLMN, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 50 …__m256i vaccOPQRSTUV = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vxOPQRSTUV, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 52 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vy01234567, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 53 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vy89ABCDEF, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 54 vaccGHIJKLMN = _mm256_add_epi32(vaccGHIJKLMN, _mm256_mullo_epi32(vyGHIJKLMN, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 55 vaccOPQRSTUV = _mm256_add_epi32(vaccOPQRSTUV, _mm256_mullo_epi32(vyOPQRSTUV, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 87 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 89 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vy01234567, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
|
D | minmax-avx2-mul32-ld64-x24.c | 45 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24() 46 …__m256i vacc89ABCDEF = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx89ABCDEF, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24() 47 …__m256i vaccGHIJKLMN = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vxGHIJKLMN, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24() 49 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vy01234567, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24() 50 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vy89ABCDEF, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24() 51 vaccGHIJKLMN = _mm256_add_epi32(vaccGHIJKLMN, _mm256_mullo_epi32(vyGHIJKLMN, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24() 81 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24() 83 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vy01234567, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x24()
|
D | minmax-avx2-mul32-ld64-x16.c | 43 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16() 44 …__m256i vacc89ABCDEF = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx89ABCDEF, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16() 46 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vy01234567, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16() 47 vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vy89ABCDEF, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16() 71 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16() 73 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vy01234567, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16()
|
D | minmax-avx2-mul32-ld64-x8.c | 41 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x8() 43 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vy01234567, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x8() 63 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x8() 65 vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vy01234567, vy_multiplier)); in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x8()
|
/external/libaom/libaom/av1/common/x86/ |
D | highbd_inv_txfm_avx2.c | 245 x = _mm256_mullo_epi32(*w0, *n0); in half_btf_0_avx2() 256 x = _mm256_mullo_epi32(*w0, *n0); in half_btf_avx2() 257 y = _mm256_mullo_epi32(*w1, *n1); in half_btf_avx2() 458 x = _mm256_mullo_epi32(in[0], cospi32); in idct32_low1_avx2() 1170 in[0] = _mm256_mullo_epi32(in[0], cospi32); in idct16_low1_avx2() 1269 x = _mm256_mullo_epi32(u[0], cospi32); in idct16_low8_avx2() 1291 x = _mm256_mullo_epi32(u[5], cospi32); in idct16_low8_avx2() 1292 y = _mm256_mullo_epi32(u[6], cospi32); in idct16_low8_avx2() 1312 x = _mm256_mullo_epi32(u[10], cospi32); in idct16_low8_avx2() 1313 y = _mm256_mullo_epi32(u[13], cospi32); in idct16_low8_avx2() [all …]
|
D | selfguided_avx2.c | 208 an = _mm256_max_epi32(_mm256_mullo_epi32(a, _mm256_set1_epi32(n)), bb); in compute_p() 211 an = _mm256_mullo_epi32(sum2, _mm256_set1_epi32(n)); in compute_p() 262 _mm256_srli_epi32(_mm256_add_epi32(_mm256_mullo_epi32(p, s), rnd_z), in calc_ab() 277 const __m256i b_int = _mm256_mullo_epi32(a_comp_over_n, sum1); in calc_ab() 399 _mm256_srli_epi32(_mm256_add_epi32(_mm256_mullo_epi32(p, s), rnd_z), in calc_ab_fast() 414 const __m256i b_int = _mm256_mullo_epi32(a_comp_over_n, sum1); in calc_ab_fast() 680 v_0 = _mm256_add_epi32(v_0, _mm256_mullo_epi32(xq0, f1_0)); in av1_apply_selfguided_restoration_avx2() 683 v_1 = _mm256_add_epi32(v_1, _mm256_mullo_epi32(xq0, f1_1)); in av1_apply_selfguided_restoration_avx2() 688 v_0 = _mm256_add_epi32(v_0, _mm256_mullo_epi32(xq1, f2_0)); in av1_apply_selfguided_restoration_avx2() 691 v_1 = _mm256_add_epi32(v_1, _mm256_mullo_epi32(xq1, f2_1)); in av1_apply_selfguided_restoration_avx2()
|
/external/XNNPACK/src/qs8-vaddc/gen/ |
D | minmax-avx2-mul32-ld64-x32.c | 43 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32() 44 …__m256i vacc89ABCDEF = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx89ABCDEF, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32() 45 …__m256i vaccGHIJKLMN = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vxGHIJKLMN, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32() 46 …__m256i vaccOPQRSTUV = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vxOPQRSTUV, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32() 76 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32()
|
D | minmax-avx2-mul32-ld64-x24.c | 42 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x24() 43 …__m256i vacc89ABCDEF = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx89ABCDEF, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x24() 44 …__m256i vaccGHIJKLMN = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vxGHIJKLMN, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x24() 72 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x24()
|
D | minmax-avx2-mul32-ld64-x16.c | 41 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x16() 42 …__m256i vacc89ABCDEF = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx89ABCDEF, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x16() 64 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x16()
|
D | minmax-avx2-mul32-ld64-x8.c | 40 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x8() 59 …__m256i vacc01234567 = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx01234567, vx_mul… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x8()
|
/external/XNNPACK/src/qs8-vadd/ |
D | avx2-mul32-ld64.c.in | 49 …__m256i vacc${ABC[N:N+8]} = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx${ABC[N:N+8… 52 …vacc${ABC[N:N+8]} = _mm256_add_epi32(vacc${ABC[N:N+8]}, _mm256_mullo_epi32(vy${ABC[N:N+8]}, vy_mul… 101 …__m256i vacc${ABC[0:8]} = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx${ABC[0:8]}, … 103 …vacc${ABC[0:8]} = _mm256_add_epi32(vacc${ABC[0:8]}, _mm256_mullo_epi32(vy${ABC[0:8]}, vy_multiplie…
|
/external/libgav1/libgav1/src/dsp/x86/ |
D | cdef_avx2.cc | 392 const __m256i c = _mm256_mullo_epi32(square_lo, division_table_0); in Cost0Or4_Pair() 393 const __m256i d = _mm256_mullo_epi32(square_hi, division_table_1); in Cost0Or4_Pair() 427 const __m256i c = _mm256_mullo_epi32(square_lo, division_table[0]); in CostOdd_Pair() 428 const __m256i d = _mm256_mullo_epi32(square_hi, division_table[1]); in CostOdd_Pair() 445 const __m256i b = _mm256_mullo_epi32(square_a, division_table); in Cost2And6_Pair()
|
/external/XNNPACK/src/qs8-vaddc/ |
D | avx2-mul32-ld64.c.in | 47 …__m256i vacc${ABC[N:N+8]} = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx${ABC[N:N+8… 94 …__m256i vacc${ABC[0:8]} = _mm256_add_epi32(vzero_point_product, _mm256_mullo_epi32(vx${ABC[0:8]}, …
|
/external/libaom/libaom/aom_dsp/x86/ |
D | txfm_common_avx2.h | 283 const __m256i r1 = _mm256_mullo_epi32(sqrt2, r0); in av1_round_shift_rect_array_32_avx2() 290 const __m256i r1 = _mm256_mullo_epi32(sqrt2, r0); in av1_round_shift_rect_array_32_avx2()
|
D | highbd_quantize_intrin_avx2.c | 82 __m256i dq = _mm256_mullo_epi32(q, qp[3]); in quantize()
|
/external/ruy/ruy/ |
D | kernel_avx2_fma.cc | 137 const __m256i rhs_sums_offset_v = _mm256_mullo_epi32( 185 const __m256i lhs_sums_offset = _mm256_mullo_epi32( 762 const __m256i rhs_sums_offset_v = _mm256_mullo_epi32( 788 const __m256i lhs_sums_offset = _mm256_mullo_epi32(
|