Home
last modified time | relevance | path

Searched refs:vb89ABc0 (Results 1 – 25 of 64) sorted by relevance

123

/external/XNNPACK/src/qs8-gemm/gen/
D4x16c2-minmax-rndnu-neon-mull-ld2r.c93 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r() local
129 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
130 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
131 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
132 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
268 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r() local
276 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
285 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
294 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
303 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
D4x16c2-minmax-rndnu-neon-mull-ld1r.c101 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r() local
137 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
138 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
139 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
140 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
276 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r() local
284 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
293 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
302 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
311 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
D4x16c2-minmax-rndnu-neon-mull-dup.c89 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup() local
125 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
126 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
127 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
128 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
264 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup() local
272 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
281 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
290 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
299 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
D4x16c2-minmax-rndnu-neon-mull-ld4r.c89 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r() local
125 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
126 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
127 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
128 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
264 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r() local
272 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
281 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
290 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
299 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
D3x16c2-minmax-rndnu-neon-mull-dup.c78 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup() local
109 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
110 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
111 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
216 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup() local
224 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
233 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
242 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
D3x16c2-minmax-rndnu-neon-mull-ld2r.c81 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r() local
112 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
113 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
114 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
219 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r() local
227 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
236 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
245 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
D3x16c2-minmax-rndnu-neon-mull-ld4r.c78 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r() local
109 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
110 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
111 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
216 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r() local
224 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
233 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
242 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
D3x16c2-minmax-rndnu-neon-mull-ld1r.c87 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r() local
118 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
119 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
120 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
225 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r() local
233 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
242 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
251 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
D2x16c2-minmax-rndnu-neon-mull-dup.c67 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup() local
93 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup()
94 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup()
168 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup() local
176 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup()
185 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup()
D2x16c2-minmax-rndnu-neon-mull-ld1r.c73 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r() local
99 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r()
100 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r()
174 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r() local
182 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r()
191 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r()
D2x16c2-minmax-rndnu-neon-mull-ld4r.c67 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r() local
93 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r()
94 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r()
168 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r() local
176 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r()
185 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r()
D2x16c2-minmax-rndnu-neon-mull-ld2r.c69 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r() local
95 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r()
96 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r()
170 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r() local
178 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r()
187 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r()
D1x16c2-minmax-rndnu-neon-mull-ld2r.c57 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c2__neon_mull_ld2r() local
78 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c2__neon_mull_ld2r()
121 const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c2__neon_mull_ld2r() local
129 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c2__neon_mull_ld2r()
/external/XNNPACK/src/qs8-igemm/gen/
D4x16c2-minmax-rndnu-neon-mull-dup.c106 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup() local
142 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
143 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
144 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
145 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
281 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup() local
289 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
298 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
307 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
316 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup()
D4x16c2-minmax-rndnu-neon-mull-ld4r.c106 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r() local
142 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
143 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
144 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
145 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
281 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r() local
289 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
298 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
307 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
316 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r()
D4x16c2-minmax-rndnu-neon-mull-ld2r.c110 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r() local
146 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
147 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
148 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
149 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
285 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r() local
293 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
302 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
311 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
320 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r()
D4x16c2-minmax-rndnu-neon-mull-ld1r.c118 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r() local
154 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
155 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
156 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
157 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
293 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r() local
301 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
310 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
319 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
328 const int16x8_t vprod3x89ABc0 = vmull_s8(vb89ABc0, va3c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r()
D3x16c2-minmax-rndnu-neon-mull-ld1r.c102 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r() local
133 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
134 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
135 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
240 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r() local
248 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
257 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
266 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld1r()
D3x16c2-minmax-rndnu-neon-mull-dup.c93 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup() local
124 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
125 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
126 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
231 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup() local
239 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
248 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
257 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_dup()
D3x16c2-minmax-rndnu-neon-mull-ld4r.c93 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r() local
124 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
125 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
126 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
231 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r() local
239 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
248 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
257 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld4r()
D3x16c2-minmax-rndnu-neon-mull-ld2r.c96 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r() local
127 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
128 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
129 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
234 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r() local
242 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
251 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
260 const int16x8_t vprod2x89ABc0 = vmull_s8(vb89ABc0, va2c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mull_ld2r()
D2x16c2-minmax-rndnu-neon-mull-ld1r.c86 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r() local
112 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r()
113 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r()
187 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r() local
195 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r()
204 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld1r()
D2x16c2-minmax-rndnu-neon-mull-dup.c80 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup() local
106 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup()
107 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup()
181 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup() local
189 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup()
198 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_dup()
D2x16c2-minmax-rndnu-neon-mull-ld4r.c80 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r() local
106 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r()
107 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r()
181 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r() local
189 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r()
198 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld4r()
D2x16c2-minmax-rndnu-neon-mull-ld2r.c82 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r() local
108 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r()
109 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r()
183 … const int8x8_t vb89ABc0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r() local
191 const int16x8_t vprod0x89ABc0 = vmull_s8(vb89ABc0, va0c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r()
200 const int16x8_t vprod1x89ABc0 = vmull_s8(vb89ABc0, va1c0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mull_ld2r()

123