Home
last modified time | relevance | path

Searched refs:vprod0x01234567c3 (Results 1 – 16 of 16) sorted by relevance

/external/XNNPACK/src/qs8-gemm/gen/
D1x8-minmax-rndnu-neon-mull-addw-dup.c67 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
68 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
69 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
119 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
120 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
121 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
D2x8-minmax-rndnu-neon-mull-addw-dup.c85 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
86 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
87 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
162 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
163 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
164 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
D1x16-minmax-rndnu-neon-mull-addw-dup.c84 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
85 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
86 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
174 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
175 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
176 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
D3x8-minmax-rndnu-neon-mull-addw-dup.c103 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
104 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
105 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
205 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
206 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
207 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
D4x8-minmax-rndnu-neon-mull-addw-dup.c121 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
122 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
123 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
248 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
249 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
250 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
D2x16-minmax-rndnu-neon-mull-addw-dup.c113 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
114 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
115 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
252 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
253 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
254 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
D3x16-minmax-rndnu-neon-mull-addw-dup.c142 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
143 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
144 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
330 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
331 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
332 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
D4x16-minmax-rndnu-neon-mull-addw-dup.c171 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
172 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
173 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
408 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
409 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
410 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); 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.c78 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
79 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
80 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
130 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
131 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
132 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
D2x8-minmax-rndnu-neon-mull-addw-dup.c98 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
99 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
100 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
175 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
176 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
177 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
D1x16-minmax-rndnu-neon-mull-addw-dup.c95 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
96 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
97 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
185 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
186 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
187 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
D3x8-minmax-rndnu-neon-mull-addw-dup.c118 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
119 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
120 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
220 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
221 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
222 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
D4x8-minmax-rndnu-neon-mull-addw-dup.c138 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
139 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
140 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
265 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
266 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
267 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
D2x16-minmax-rndnu-neon-mull-addw-dup.c126 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
127 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
128 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
265 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
266 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
267 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
D3x16-minmax-rndnu-neon-mull-addw-dup.c157 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
158 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
159 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
345 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
346 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
347 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
D4x16-minmax-rndnu-neon-mull-addw-dup.c188 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
189 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
190 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
425 const int16x8_t vprod0x01234567c3 = vmull_s8(vb01234567c3, vdup_lane_s8(va0, 3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
426 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
427 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c3)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()