Home
last modified time | relevance | path

Searched refs:_mm_mullo_epi16 (Results 1 – 25 of 105) sorted by relevance

12345

/external/libavc/common/x86/
Dih264_weighted_pred_sse42.c419 y_0_8x16b = _mm_mullo_epi16(y_0_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
420 y_2_8x16b = _mm_mullo_epi16(y_2_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
463 y_0_8x16b = _mm_mullo_epi16(y_0_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
464 y_1_8x16b = _mm_mullo_epi16(y_1_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
465 y_2_8x16b = _mm_mullo_epi16(y_2_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
466 y_3_8x16b = _mm_mullo_epi16(y_3_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
523 y_0L_8x16b = _mm_mullo_epi16(y_0L_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
524 y_0H_8x16b = _mm_mullo_epi16(y_0H_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
525 y_1L_8x16b = _mm_mullo_epi16(y_1L_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
526 y_1H_8x16b = _mm_mullo_epi16(y_1H_8x16b, wt_8x16b); in ih264_weighted_pred_luma_sse42()
[all …]
/external/XNNPACK/src/qs8-vadd/gen/
Dminmax-sse41-mul16-ld64-x32.c51 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
52 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
55 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
56 const __m128i vyprod89ABCDEFlo = _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
59 const __m128i vxprodGHIJKLMNlo = _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
60 const __m128i vyprodGHIJKLMNlo = _mm_mullo_epi16(vyGHIJKLMN, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
63 const __m128i vxprodOPQRSTUVlo = _mm_mullo_epi16(vxOPQRSTUV, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
64 const __m128i vyprodOPQRSTUVlo = _mm_mullo_epi16(vyOPQRSTUV, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
66 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
67 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x32()
[all …]
Dminmax-sse2-mul16-ld64-x32.c59 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
60 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
63 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
64 const __m128i vyprod89ABCDEFlo = _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
67 const __m128i vxprodGHIJKLMNlo = _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
68 const __m128i vyprodGHIJKLMNlo = _mm_mullo_epi16(vyGHIJKLMN, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
71 const __m128i vxprodOPQRSTUVlo = _mm_mullo_epi16(vxOPQRSTUV, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
72 const __m128i vyprodOPQRSTUVlo = _mm_mullo_epi16(vyOPQRSTUV, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
74 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
75 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x32()
[all …]
Dminmax-sse41-mul16-ld64-x24.c49 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
50 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
53 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
54 const __m128i vyprod89ABCDEFlo = _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
57 const __m128i vxprodGHIJKLMNlo = _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
58 const __m128i vyprodGHIJKLMNlo = _mm_mullo_epi16(vyGHIJKLMN, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
60 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
61 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
62 … vxprod89ABCDEFhi = _mm_add_epi16(vxprod89ABCDEFhi, _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
63 … vyprod89ABCDEFhi = _mm_add_epi16(vyprod89ABCDEFhi, _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x24()
[all …]
Dminmax-sse2-mul16-ld64-x24.c55 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
56 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
59 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
60 const __m128i vyprod89ABCDEFlo = _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
63 const __m128i vxprodGHIJKLMNlo = _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
64 const __m128i vyprodGHIJKLMNlo = _mm_mullo_epi16(vyGHIJKLMN, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
66 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
67 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
68 … vxprod89ABCDEFhi = _mm_add_epi16(vxprod89ABCDEFhi, _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
69 … vyprod89ABCDEFhi = _mm_add_epi16(vyprod89ABCDEFhi, _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x24()
[all …]
Dminmax-sse2-mul16-ld64-x16.c51 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
52 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
55 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
56 const __m128i vyprod89ABCDEFlo = _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
58 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
59 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
60 … vxprod89ABCDEFhi = _mm_add_epi16(vxprod89ABCDEFhi, _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
61 … vyprod89ABCDEFhi = _mm_add_epi16(vyprod89ABCDEFhi, _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
114 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
115 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x16()
[all …]
Dminmax-sse41-mul16-ld64-x16.c47 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
48 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
51 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
52 const __m128i vyprod89ABCDEFlo = _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
54 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
55 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
56 … vxprod89ABCDEFhi = _mm_add_epi16(vxprod89ABCDEFhi, _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
57 … vyprod89ABCDEFhi = _mm_add_epi16(vyprod89ABCDEFhi, _mm_mullo_epi16(vy89ABCDEF, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
108 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
109 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x16()
[all …]
Dminmax-sse41-mul16-ld64-x8.c45 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x8()
46 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x8()
48 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x8()
49 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x8()
85 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x8()
86 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x8()
88 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x8()
89 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse41_mul16_ld64_x8()
Dminmax-sse2-mul16-ld64-x8.c47 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x8()
48 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x8()
50 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x8()
51 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x8()
89 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x8()
90 const __m128i vyprod01234567lo = _mm_mullo_epi16(vy01234567, vy_multiplier_lo); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x8()
92 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x8()
93 … vyprod01234567hi = _mm_add_epi16(vyprod01234567hi, _mm_mullo_epi16(vy01234567, vy_multiplier_hi)); in xnn_qs8_vadd_minmax_ukernel__sse2_mul16_ld64_x8()
/external/libvpx/libvpx/vp8/common/x86/
Dbilinear_filter_sse2.c47 const __m128i a_lo_filtered = _mm_mullo_epi16(a_lo, hfilter_0); in horizontal_16x16()
48 const __m128i a_hi_filtered = _mm_mullo_epi16(a_hi, hfilter_0); in horizontal_16x16()
53 const __m128i b_lo_filtered = _mm_mullo_epi16(b_lo, hfilter_1); in horizontal_16x16()
54 const __m128i b_hi_filtered = _mm_mullo_epi16(b_hi, hfilter_1); in horizontal_16x16()
100 const __m128i row_0_lo_filtered = _mm_mullo_epi16(row_0_lo, vfilter_0); in vertical_16x16()
101 const __m128i row_0_hi_filtered = _mm_mullo_epi16(row_0_hi, vfilter_0); in vertical_16x16()
105 const __m128i row_1_lo_filtered = _mm_mullo_epi16(row_1_lo, vfilter_1); in vertical_16x16()
106 const __m128i row_1_hi_filtered = _mm_mullo_epi16(row_1_hi, vfilter_1); in vertical_16x16()
172 const __m128i a_filtered = _mm_mullo_epi16(a_u16, hfilter_0); in horizontal_8xN()
173 const __m128i b_filtered = _mm_mullo_epi16(b_u16, hfilter_1); in horizontal_8xN()
[all …]
/external/XNNPACK/src/qs8-dwconv/gen/
Dup24x9-minmax-sse41-mul16.c106 const __m128i vp0x01234567lo = _mm_mullo_epi16(vxi0x01234567, vxk0x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
108 const __m128i vp0x89ABCDEFlo = _mm_mullo_epi16(vxi0x89ABCDEF, vxk0x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
110 const __m128i vp0xGHIJKLMNlo = _mm_mullo_epi16(vxi0xGHIJKLMN, vxk0xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
135 const __m128i vp1x01234567lo = _mm_mullo_epi16(vxi1x01234567, vxk1x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
137 const __m128i vp1x89ABCDEFlo = _mm_mullo_epi16(vxi1x89ABCDEF, vxk1x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
139 const __m128i vp1xGHIJKLMNlo = _mm_mullo_epi16(vxi1xGHIJKLMN, vxk1xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
164 const __m128i vp2x01234567lo = _mm_mullo_epi16(vxi2x01234567, vxk2x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
166 const __m128i vp2x89ABCDEFlo = _mm_mullo_epi16(vxi2x89ABCDEF, vxk2x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
168 const __m128i vp2xGHIJKLMNlo = _mm_mullo_epi16(vxi2xGHIJKLMN, vxk2xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
193 const __m128i vp3x01234567lo = _mm_mullo_epi16(vxi3x01234567, vxk3x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
[all …]
Dup16x9-minmax-sse41-mul16.c100 const __m128i vp0x01234567lo = _mm_mullo_epi16(vxi0x01234567, vxk0x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
102 const __m128i vp0x89ABCDEFlo = _mm_mullo_epi16(vxi0x89ABCDEF, vxk0x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
121 const __m128i vp1x01234567lo = _mm_mullo_epi16(vxi1x01234567, vxk1x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
123 const __m128i vp1x89ABCDEFlo = _mm_mullo_epi16(vxi1x89ABCDEF, vxk1x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
142 const __m128i vp2x01234567lo = _mm_mullo_epi16(vxi2x01234567, vxk2x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
144 const __m128i vp2x89ABCDEFlo = _mm_mullo_epi16(vxi2x89ABCDEF, vxk2x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
163 const __m128i vp3x01234567lo = _mm_mullo_epi16(vxi3x01234567, vxk3x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
165 const __m128i vp3x89ABCDEFlo = _mm_mullo_epi16(vxi3x89ABCDEF, vxk3x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
184 const __m128i vp4x01234567lo = _mm_mullo_epi16(vxi4x01234567, vxk4x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
186 const __m128i vp4x89ABCDEFlo = _mm_mullo_epi16(vxi4x89ABCDEF, vxk4x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse41_mul16()
[all …]
Dup8x9-minmax-sse41-mul16.c94 const __m128i vp0x01234567lo = _mm_mullo_epi16(vxi0x01234567, vxk0x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
107 const __m128i vp1x01234567lo = _mm_mullo_epi16(vxi1x01234567, vxk1x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
120 const __m128i vp2x01234567lo = _mm_mullo_epi16(vxi2x01234567, vxk2x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
133 const __m128i vp3x01234567lo = _mm_mullo_epi16(vxi3x01234567, vxk3x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
146 const __m128i vp4x01234567lo = _mm_mullo_epi16(vxi4x01234567, vxk4x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
159 const __m128i vp5x01234567lo = _mm_mullo_epi16(vxi5x01234567, vxk5x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
172 const __m128i vp6x01234567lo = _mm_mullo_epi16(vxi6x01234567, vxk6x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
185 const __m128i vp7x01234567lo = _mm_mullo_epi16(vxi7x01234567, vxk7x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
198 const __m128i vp8x01234567lo = _mm_mullo_epi16(vxi8x01234567, vxk8x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
261 const __m128i vp0x01234567lo = _mm_mullo_epi16(vxi0x01234567, vxk0x01234567); in xnn_qs8_dwconv_minmax_ukernel_up8x9__sse41_mul16()
[all …]
Dup16x9-minmax-ssse3-mul16.c100 const __m128i vp0x01234567lo = _mm_mullo_epi16(vxi0x01234567, vxk0x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
102 const __m128i vp0x89ABCDEFlo = _mm_mullo_epi16(vxi0x89ABCDEF, vxk0x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
121 const __m128i vp1x01234567lo = _mm_mullo_epi16(vxi1x01234567, vxk1x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
123 const __m128i vp1x89ABCDEFlo = _mm_mullo_epi16(vxi1x89ABCDEF, vxk1x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
142 const __m128i vp2x01234567lo = _mm_mullo_epi16(vxi2x01234567, vxk2x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
144 const __m128i vp2x89ABCDEFlo = _mm_mullo_epi16(vxi2x89ABCDEF, vxk2x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
163 const __m128i vp3x01234567lo = _mm_mullo_epi16(vxi3x01234567, vxk3x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
165 const __m128i vp3x89ABCDEFlo = _mm_mullo_epi16(vxi3x89ABCDEF, vxk3x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
184 const __m128i vp4x01234567lo = _mm_mullo_epi16(vxi4x01234567, vxk4x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
186 const __m128i vp4x89ABCDEFlo = _mm_mullo_epi16(vxi4x89ABCDEF, vxk4x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__ssse3_mul16()
[all …]
Dup16x9-minmax-sse2-mul16.c100 const __m128i vp0x01234567lo = _mm_mullo_epi16(vxi0x01234567, vxk0x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
102 const __m128i vp0x89ABCDEFlo = _mm_mullo_epi16(vxi0x89ABCDEF, vxk0x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
121 const __m128i vp1x01234567lo = _mm_mullo_epi16(vxi1x01234567, vxk1x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
123 const __m128i vp1x89ABCDEFlo = _mm_mullo_epi16(vxi1x89ABCDEF, vxk1x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
142 const __m128i vp2x01234567lo = _mm_mullo_epi16(vxi2x01234567, vxk2x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
144 const __m128i vp2x89ABCDEFlo = _mm_mullo_epi16(vxi2x89ABCDEF, vxk2x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
163 const __m128i vp3x01234567lo = _mm_mullo_epi16(vxi3x01234567, vxk3x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
165 const __m128i vp3x89ABCDEFlo = _mm_mullo_epi16(vxi3x89ABCDEF, vxk3x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
184 const __m128i vp4x01234567lo = _mm_mullo_epi16(vxi4x01234567, vxk4x01234567); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
186 const __m128i vp4x89ABCDEFlo = _mm_mullo_epi16(vxi4x89ABCDEF, vxk4x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up16x9__sse2_mul16()
[all …]
Dup24x9-minmax-ssse3-mul16.c106 const __m128i vp0x01234567lo = _mm_mullo_epi16(vxi0x01234567, vxk0x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
108 const __m128i vp0x89ABCDEFlo = _mm_mullo_epi16(vxi0x89ABCDEF, vxk0x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
110 const __m128i vp0xGHIJKLMNlo = _mm_mullo_epi16(vxi0xGHIJKLMN, vxk0xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
135 const __m128i vp1x01234567lo = _mm_mullo_epi16(vxi1x01234567, vxk1x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
137 const __m128i vp1x89ABCDEFlo = _mm_mullo_epi16(vxi1x89ABCDEF, vxk1x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
139 const __m128i vp1xGHIJKLMNlo = _mm_mullo_epi16(vxi1xGHIJKLMN, vxk1xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
164 const __m128i vp2x01234567lo = _mm_mullo_epi16(vxi2x01234567, vxk2x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
166 const __m128i vp2x89ABCDEFlo = _mm_mullo_epi16(vxi2x89ABCDEF, vxk2x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
168 const __m128i vp2xGHIJKLMNlo = _mm_mullo_epi16(vxi2xGHIJKLMN, vxk2xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
193 const __m128i vp3x01234567lo = _mm_mullo_epi16(vxi3x01234567, vxk3x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
[all …]
Dup24x9-minmax-sse2-mul16.c106 const __m128i vp0x01234567lo = _mm_mullo_epi16(vxi0x01234567, vxk0x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
108 const __m128i vp0x89ABCDEFlo = _mm_mullo_epi16(vxi0x89ABCDEF, vxk0x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
110 const __m128i vp0xGHIJKLMNlo = _mm_mullo_epi16(vxi0xGHIJKLMN, vxk0xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
135 const __m128i vp1x01234567lo = _mm_mullo_epi16(vxi1x01234567, vxk1x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
137 const __m128i vp1x89ABCDEFlo = _mm_mullo_epi16(vxi1x89ABCDEF, vxk1x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
139 const __m128i vp1xGHIJKLMNlo = _mm_mullo_epi16(vxi1xGHIJKLMN, vxk1xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
164 const __m128i vp2x01234567lo = _mm_mullo_epi16(vxi2x01234567, vxk2x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
166 const __m128i vp2x89ABCDEFlo = _mm_mullo_epi16(vxi2x89ABCDEF, vxk2x89ABCDEF); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
168 const __m128i vp2xGHIJKLMNlo = _mm_mullo_epi16(vxi2xGHIJKLMN, vxk2xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
193 const __m128i vp3x01234567lo = _mm_mullo_epi16(vxi3x01234567, vxk3x01234567); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
[all …]
/external/libgav1/libgav1/src/dsp/x86/
Dintrapred_smooth_sse4.cc62 const __m128i weighted_left_y = _mm_mullo_epi16(left_y, weights); in WriteSmoothHorizontalSum4()
76 const __m128i weighted_px = _mm_mullo_epi16(pixels, weights); in SmoothDirectionalSum8()
101 const __m128i weighted_px1 = _mm_mullo_epi16(pixels1, weights1); in WriteSmoothDirectionalSum16()
102 const __m128i weighted_px2 = _mm_mullo_epi16(pixels2, weights2); in WriteSmoothDirectionalSum16()
229 const __m128i scaled_top_right = _mm_mullo_epi16(inverted_weights, top_right); in Smooth4x4_SSE4_1()
231 _mm_mullo_epi16(inverted_weights, bottom_left); in Smooth4x4_SSE4_1()
493 _mm_mullo_epi16(scale_m_weights_y, bottom_left); in SmoothWxH()
515 _mm_mullo_epi16(inverted_weights_x, top_right); in SmoothWxH()
544 const __m128i scaled_top_right = _mm_mullo_epi16(inverted_weights, top_right); in SmoothHorizontal4x4_SSE4_1()
564 const __m128i scaled_top_right = _mm_mullo_epi16(inverted_weights, top_right); in SmoothHorizontal4x8_SSE4_1()
[all …]
/external/XNNPACK/src/qu8-vadd/
Dminmax-sse2.c43 const __m128i va_product_lo = _mm_mullo_epi16(vxa, va_multiplier_lo); in xnn_qu8_vadd_minmax_ukernel__sse2()
45 _mm_add_epi16(_mm_mulhi_epu16(vxa, va_multiplier_lo), _mm_mullo_epi16(vxa, va_multiplier_hi)); in xnn_qu8_vadd_minmax_ukernel__sse2()
47 const __m128i vb_product_lo = _mm_mullo_epi16(vxb, vb_multiplier_lo); in xnn_qu8_vadd_minmax_ukernel__sse2()
49 _mm_add_epi16(_mm_mulhi_epu16(vxb, vb_multiplier_lo), _mm_mullo_epi16(vxb, vb_multiplier_hi)); in xnn_qu8_vadd_minmax_ukernel__sse2()
85 const __m128i va_product_lo = _mm_mullo_epi16(vxa, va_multiplier_lo); in xnn_qu8_vadd_minmax_ukernel__sse2()
87 _mm_add_epi16(_mm_mulhi_epu16(vxa, va_multiplier_lo), _mm_mullo_epi16(vxa, va_multiplier_hi)); in xnn_qu8_vadd_minmax_ukernel__sse2()
89 const __m128i vb_product_lo = _mm_mullo_epi16(vxb, vb_multiplier_lo); in xnn_qu8_vadd_minmax_ukernel__sse2()
91 _mm_add_epi16(_mm_mulhi_epu16(vxb, vb_multiplier_lo), _mm_mullo_epi16(vxb, vb_multiplier_hi)); in xnn_qu8_vadd_minmax_ukernel__sse2()
/external/XNNPACK/src/qs8-vaddc/gen/
Dminmax-sse2-mul16-ld64-x32.c49 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
51 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
53 const __m128i vxprodGHIJKLMNlo = _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
55 const __m128i vxprodOPQRSTUVlo = _mm_mullo_epi16(vxOPQRSTUV, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
57 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
58 … vxprod89ABCDEFhi = _mm_add_epi16(vxprod89ABCDEFhi, _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
59 … vxprodGHIJKLMNhi = _mm_add_epi16(vxprodGHIJKLMNhi, _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
60 … vxprodOPQRSTUVhi = _mm_add_epi16(vxprodOPQRSTUVhi, _mm_mullo_epi16(vxOPQRSTUV, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
124 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
126 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x32()
Dminmax-sse41-mul16-ld64-x32.c45 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
47 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
49 const __m128i vxprodGHIJKLMNlo = _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
51 const __m128i vxprodOPQRSTUVlo = _mm_mullo_epi16(vxOPQRSTUV, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
53 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
54 … vxprod89ABCDEFhi = _mm_add_epi16(vxprod89ABCDEFhi, _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
55 … vxprodGHIJKLMNhi = _mm_add_epi16(vxprodGHIJKLMNhi, _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
56 … vxprodOPQRSTUVhi = _mm_add_epi16(vxprodOPQRSTUVhi, _mm_mullo_epi16(vxOPQRSTUV, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
119 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
121 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x32()
Dminmax-sse41-mul16-ld64-x24.c44 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x24()
46 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x24()
48 const __m128i vxprodGHIJKLMNlo = _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x24()
50 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x24()
51 … vxprod89ABCDEFhi = _mm_add_epi16(vxprod89ABCDEFhi, _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x24()
52 … vxprodGHIJKLMNhi = _mm_add_epi16(vxprodGHIJKLMNhi, _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x24()
105 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x24()
107 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse41_mul16_ld64_x24()
Dminmax-sse2-mul16-ld64-x24.c47 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x24()
49 const __m128i vxprod89ABCDEFlo = _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x24()
51 const __m128i vxprodGHIJKLMNlo = _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x24()
53 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x24()
54 … vxprod89ABCDEFhi = _mm_add_epi16(vxprod89ABCDEFhi, _mm_mullo_epi16(vx89ABCDEF, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x24()
55 … vxprodGHIJKLMNhi = _mm_add_epi16(vxprodGHIJKLMNhi, _mm_mullo_epi16(vxGHIJKLMN, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x24()
109 const __m128i vxprod01234567lo = _mm_mullo_epi16(vx01234567, vx_multiplier_lo); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x24()
111 … vxprod01234567hi = _mm_add_epi16(vxprod01234567hi, _mm_mullo_epi16(vx01234567, vx_multiplier_hi)); in xnn_qs8_vaddc_minmax_ukernel__sse2_mul16_ld64_x24()
/external/libaom/libaom/aom_dsp/x86/
Dblend_sse4.h33 const __m128i v_p0_w = _mm_mullo_epi16(v_s0_w, *v_m0_w); in blend_4()
34 const __m128i v_p1_w = _mm_mullo_epi16(v_s1_w, *v_m1_w); in blend_4()
48 const __m128i v_p0_w = _mm_mullo_epi16(v_s0_w, *v_m0_w); in blend_8()
49 const __m128i v_p1_w = _mm_mullo_epi16(v_s1_w, *v_m1_w); in blend_8()
111 const __m128i v_p0_w = _mm_mullo_epi16(v_s0_w, v_m0_w); in blend_4_b10()
112 const __m128i v_p1_w = _mm_mullo_epi16(v_s1_w, v_m1_w); in blend_4_b10()
126 const __m128i v_p0_w = _mm_mullo_epi16(v_s0_w, v_m0_w); in blend_8_b10()
127 const __m128i v_p1_w = _mm_mullo_epi16(v_s1_w, v_m1_w); in blend_8_b10()
/external/XNNPACK/src/qu8-dwconv/
Dup8x9-minmax-sse2.c79 const __m128i vprod0_odd = _mm_mullo_epi16(vxi0, vxk0); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
88 const __m128i vprod1_odd = _mm_mullo_epi16(vxi1, vxk1); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
97 const __m128i vprod2_odd = _mm_mullo_epi16(vxi2, vxk2); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
106 const __m128i vprod3_odd = _mm_mullo_epi16(vxi3, vxk3); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
115 const __m128i vprod4_odd = _mm_mullo_epi16(vxi4, vxk4); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
124 const __m128i vprod5_odd = _mm_mullo_epi16(vxi5, vxk5); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
133 const __m128i vprod6_odd = _mm_mullo_epi16(vxi6, vxk6); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
142 const __m128i vprod7_odd = _mm_mullo_epi16(vxi7, vxk7); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
151 const __m128i vprod8_odd = _mm_mullo_epi16(vxi8, vxk8); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
231 const __m128i vprod0_odd = _mm_mullo_epi16(vxi0, vxk0); in xnn_qu8_dwconv_minmax_ukernel_up8x9__sse2()
[all …]

12345