Lines Matching refs:vdup_lane_s16
109 …int16x8_t vprod0x0123c0 = vmull_s8(vb0123c0x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
110 …int16x8_t vprod1x0123c0 = vmull_s8(vb0123c0x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
111 …int16x8_t vprod2x0123c0 = vmull_s8(vb0123c0x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
112 …int16x8_t vprod3x0123c0 = vmull_s8(vb0123c0x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
114 …vprod0x0123c0 = vmlal_s8(vprod0x0123c0, vb0123c0x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
115 …vprod1x0123c0 = vmlal_s8(vprod1x0123c0, vb0123c0x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
116 …vprod2x0123c0 = vmlal_s8(vprod2x0123c0, vb0123c0x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
117 …vprod3x0123c0 = vmlal_s8(vprod3x0123c0, vb0123c0x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
122 …int16x8_t vprod0x4567c0 = vmull_s8(vb4567c0x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
123 …int16x8_t vprod1x4567c0 = vmull_s8(vb4567c0x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
124 …int16x8_t vprod2x4567c0 = vmull_s8(vb4567c0x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
125 …int16x8_t vprod3x4567c0 = vmull_s8(vb4567c0x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
127 …vprod0x4567c0 = vmlal_s8(vprod0x4567c0, vb4567c0x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
128 …vprod1x4567c0 = vmlal_s8(vprod1x4567c0, vb4567c0x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
129 …vprod2x4567c0 = vmlal_s8(vprod2x4567c0, vb4567c0x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
130 …vprod3x4567c0 = vmlal_s8(vprod3x4567c0, vb4567c0x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
135 …int16x8_t vprod0x0123c1 = vmull_s8(vb0123c1x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
136 …int16x8_t vprod1x0123c1 = vmull_s8(vb0123c1x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
137 …int16x8_t vprod2x0123c1 = vmull_s8(vb0123c1x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
138 …int16x8_t vprod3x0123c1 = vmull_s8(vb0123c1x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
140 …vprod0x0123c1 = vmlal_s8(vprod0x0123c1, vb0123c1x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
141 …vprod1x0123c1 = vmlal_s8(vprod1x0123c1, vb0123c1x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
142 …vprod2x0123c1 = vmlal_s8(vprod2x0123c1, vb0123c1x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
143 …vprod3x0123c1 = vmlal_s8(vprod3x0123c1, vb0123c1x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
148 …int16x8_t vprod0x4567c1 = vmull_s8(vb4567c1x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
149 …int16x8_t vprod1x4567c1 = vmull_s8(vb4567c1x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
150 …int16x8_t vprod2x4567c1 = vmull_s8(vb4567c1x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
151 …int16x8_t vprod3x4567c1 = vmull_s8(vb4567c1x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
153 …vprod0x4567c1 = vmlal_s8(vprod0x4567c1, vb4567c1x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
154 …vprod1x4567c1 = vmlal_s8(vprod1x4567c1, vb4567c1x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
155 …vprod2x4567c1 = vmlal_s8(vprod2x4567c1, vb4567c1x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
156 …vprod3x4567c1 = vmlal_s8(vprod3x4567c1, vb4567c1x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
161 …int16x8_t vprod0x0123c2 = vmull_s8(vb0123c2x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
162 …int16x8_t vprod1x0123c2 = vmull_s8(vb0123c2x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
163 …int16x8_t vprod2x0123c2 = vmull_s8(vb0123c2x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
164 …int16x8_t vprod3x0123c2 = vmull_s8(vb0123c2x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
166 …vprod0x0123c2 = vmlal_s8(vprod0x0123c2, vb0123c2x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
167 …vprod1x0123c2 = vmlal_s8(vprod1x0123c2, vb0123c2x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
168 …vprod2x0123c2 = vmlal_s8(vprod2x0123c2, vb0123c2x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
169 …vprod3x0123c2 = vmlal_s8(vprod3x0123c2, vb0123c2x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
174 …int16x8_t vprod0x4567c2 = vmull_s8(vb4567c2x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
175 …int16x8_t vprod1x4567c2 = vmull_s8(vb4567c2x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
176 …int16x8_t vprod2x4567c2 = vmull_s8(vb4567c2x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
177 …int16x8_t vprod3x4567c2 = vmull_s8(vb4567c2x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
179 …vprod0x4567c2 = vmlal_s8(vprod0x4567c2, vb4567c2x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
180 …vprod1x4567c2 = vmlal_s8(vprod1x4567c2, vb4567c2x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
181 …vprod2x4567c2 = vmlal_s8(vprod2x4567c2, vb4567c2x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
182 …vprod3x4567c2 = vmlal_s8(vprod3x4567c2, vb4567c2x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
187 …int16x8_t vprod0x0123c3 = vmull_s8(vb0123c3x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
188 …int16x8_t vprod1x0123c3 = vmull_s8(vb0123c3x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
189 …int16x8_t vprod2x0123c3 = vmull_s8(vb0123c3x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
190 …int16x8_t vprod3x0123c3 = vmull_s8(vb0123c3x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
192 …vprod0x0123c3 = vmlal_s8(vprod0x0123c3, vb0123c3x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
193 …vprod1x0123c3 = vmlal_s8(vprod1x0123c3, vb0123c3x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
194 …vprod2x0123c3 = vmlal_s8(vprod2x0123c3, vb0123c3x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
195 …vprod3x0123c3 = vmlal_s8(vprod3x0123c3, vb0123c3x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
200 …int16x8_t vprod0x4567c3 = vmull_s8(vb4567c3x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
201 …int16x8_t vprod1x4567c3 = vmull_s8(vb4567c3x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
202 …int16x8_t vprod2x4567c3 = vmull_s8(vb4567c3x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
203 …int16x8_t vprod3x4567c3 = vmull_s8(vb4567c3x0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
205 …vprod0x4567c3 = vmlal_s8(vprod0x4567c3, vb4567c3x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
206 …vprod1x4567c3 = vmlal_s8(vprod1x4567c3, vb4567c3x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
207 …vprod2x4567c3 = vmlal_s8(vprod2x4567c3, vb4567c3x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
208 …vprod3x4567c3 = vmlal_s8(vprod3x4567c3, vb4567c3x1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
232 …const int16x8_t vprod0x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
233 …const int16x8_t vprod0x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
234 …const int16x8_t vprod0x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
235 …const int16x8_t vprod0x0123c3 = vmull_s8(vb0123c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
240 …const int16x8_t vprod0x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
241 …const int16x8_t vprod0x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
242 …const int16x8_t vprod0x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
243 …const int16x8_t vprod0x4567c3 = vmull_s8(vb4567c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
248 …const int16x8_t vprod1x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
249 …const int16x8_t vprod1x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
250 …const int16x8_t vprod1x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
251 …const int16x8_t vprod1x0123c3 = vmull_s8(vb0123c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
256 …const int16x8_t vprod1x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
257 …const int16x8_t vprod1x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
258 …const int16x8_t vprod1x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
259 …const int16x8_t vprod1x4567c3 = vmull_s8(vb4567c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
264 …const int16x8_t vprod2x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
265 …const int16x8_t vprod2x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
266 …const int16x8_t vprod2x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
267 …const int16x8_t vprod2x0123c3 = vmull_s8(vb0123c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
272 …const int16x8_t vprod2x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
273 …const int16x8_t vprod2x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
274 …const int16x8_t vprod2x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
275 …const int16x8_t vprod2x4567c3 = vmull_s8(vb4567c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
280 …const int16x8_t vprod3x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
281 …const int16x8_t vprod3x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
282 …const int16x8_t vprod3x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
283 …const int16x8_t vprod3x0123c3 = vmull_s8(vb0123c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
288 …const int16x8_t vprod3x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
289 …const int16x8_t vprod3x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
290 …const int16x8_t vprod3x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
291 …const int16x8_t vprod3x4567c3 = vmull_s8(vb4567c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
309 …const int16x8_t vprod0x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
311 …const int16x8_t vprod0x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
313 …const int16x8_t vprod1x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
315 …const int16x8_t vprod1x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
317 …const int16x8_t vprod2x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
319 …const int16x8_t vprod2x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
321 …const int16x8_t vprod3x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
323 …const int16x8_t vprod3x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
330 …const int16x8_t vprod0x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
332 …const int16x8_t vprod0x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
334 …const int16x8_t vprod1x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
336 …const int16x8_t vprod1x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
338 …const int16x8_t vprod2x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
340 …const int16x8_t vprod2x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
342 …const int16x8_t vprod3x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
344 …const int16x8_t vprod3x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
351 …const int16x8_t vprod0x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
353 …const int16x8_t vprod0x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
355 …const int16x8_t vprod1x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
357 …const int16x8_t vprod1x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
359 …const int16x8_t vprod2x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
361 …const int16x8_t vprod2x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
363 …const int16x8_t vprod3x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()
365 …const int16x8_t vprod3x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_… in xnn_qs8_igemm_minmax_ukernel_4x8c2__neon_mlal_padal_dup()