Home
last modified time | relevance | path

Searched refs:vprod0x01234567c2 (Results 1 – 16 of 16) sorted by relevance

/external/XNNPACK/src/qs8-gemm/gen/
D1x8-minmax-rndnu-neon-mull-addw-dup.c62 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
63 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
64 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
112 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
113 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
114 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
D2x8-minmax-rndnu-neon-mull-addw-dup.c77 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
78 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
79 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
152 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
153 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
154 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
D1x16-minmax-rndnu-neon-mull-addw-dup.c74 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
75 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
76 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
163 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
164 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
165 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
D3x8-minmax-rndnu-neon-mull-addw-dup.c92 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
93 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
94 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
192 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
193 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
194 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
D4x8-minmax-rndnu-neon-mull-addw-dup.c107 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
108 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
109 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
232 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
233 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
234 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
D2x16-minmax-rndnu-neon-mull-addw-dup.c97 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
98 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
99 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
235 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
236 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
237 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
D3x16-minmax-rndnu-neon-mull-addw-dup.c120 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
121 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
122 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
307 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
308 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
309 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
D4x16-minmax-rndnu-neon-mull-addw-dup.c143 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
144 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
145 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
379 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
380 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
381 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); 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.c73 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
74 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
75 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
123 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup() local
124 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
125 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8__neon_mull_addw_dup()
D2x8-minmax-rndnu-neon-mull-addw-dup.c90 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
91 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
92 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
165 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup() local
166 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
167 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8__neon_mull_addw_dup()
D1x16-minmax-rndnu-neon-mull-addw-dup.c85 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
86 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
87 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
174 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup() local
175 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
176 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16__neon_mull_addw_dup()
D3x8-minmax-rndnu-neon-mull-addw-dup.c107 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
108 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
109 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
207 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup() local
208 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
209 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8__neon_mull_addw_dup()
D4x8-minmax-rndnu-neon-mull-addw-dup.c124 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
125 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
126 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
249 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup() local
250 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
251 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8__neon_mull_addw_dup()
D2x16-minmax-rndnu-neon-mull-addw-dup.c110 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
111 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
112 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
248 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup() local
249 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
250 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16__neon_mull_addw_dup()
D3x16-minmax-rndnu-neon-mull-addw-dup.c135 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
136 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
137 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
322 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup() local
323 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
324 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16__neon_mull_addw_dup()
D4x16-minmax-rndnu-neon-mull-addw-dup.c160 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
161 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
162 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
396 const int16x8_t vprod0x01234567c2 = vmull_s8(vb01234567c2, vdup_lane_s8(va0, 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup() local
397 vacc0x0123 = vaddw_s16(vacc0x0123, vget_low_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()
398 vacc0x4567 = vaddw_s16(vacc0x4567, vget_high_s16(vprod0x01234567c2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16__neon_mull_addw_dup()