Home
last modified time | relevance | path

Searched refs:vprod0x01234567c1 (Results 1 – 16 of 16) sorted by relevance

/external/XNNPACK/src/qs8-gemm/gen/
D1x8-minmax-rndnu-neon-mull-addw-dup.c57 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
58 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
59 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
105 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
106 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
107 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
D2x8-minmax-rndnu-neon-mull-addw-dup.c69 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
70 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
71 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
142 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
143 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
144 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
D1x16-minmax-rndnu-neon-mull-addw-dup.c64 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
65 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
66 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
152 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
153 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
154 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
D3x8-minmax-rndnu-neon-mull-addw-dup.c81 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
82 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
83 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
179 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
180 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
181 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
D4x8-minmax-rndnu-neon-mull-addw-dup.c93 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
94 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
95 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
216 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
217 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
218 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
D2x16-minmax-rndnu-neon-mull-addw-dup.c81 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
82 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
83 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
218 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
219 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
220 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
D3x16-minmax-rndnu-neon-mull-addw-dup.c98 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
99 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
100 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
284 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
285 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
286 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
D4x16-minmax-rndnu-neon-mull-addw-dup.c115 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
116 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
117 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
350 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
351 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
352 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
/external/XNNPACK/src/qs8-igemm/gen/
D1x8-minmax-rndnu-neon-mull-addw-dup.c68 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
69 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
70 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
116 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
117 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
118 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
D2x8-minmax-rndnu-neon-mull-addw-dup.c82 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
83 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
84 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
155 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
156 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
157 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
D1x16-minmax-rndnu-neon-mull-addw-dup.c75 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
76 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
77 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
163 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
164 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
165 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
D3x8-minmax-rndnu-neon-mull-addw-dup.c96 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
97 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
98 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
194 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
195 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
196 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
D4x8-minmax-rndnu-neon-mull-addw-dup.c110 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
111 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
112 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
233 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
234 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
235 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
D2x16-minmax-rndnu-neon-mull-addw-dup.c94 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
95 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
96 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
231 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
232 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
233 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
D3x16-minmax-rndnu-neon-mull-addw-dup.c113 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
114 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
115 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
299 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
300 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
301 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
D4x16-minmax-rndnu-neon-mull-addw-dup.c132 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
133 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
134 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
367 const int16x8_t vprod0x01234567c1 = vmull_s8(vb01234567c1, vdup_lane_s8(va0, 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
368 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
369 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()