/external/XNNPACK/src/qs8-gemm/gen/ |
D | 3x8c4s2-minmax-rndnu-neon-mlal.c | 135 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4s2__neon_mlal() local 219 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4s2__neon_mlal() local
|
D | 4x8c4s2-minmax-rndnu-neon-mlal.c | 161 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local 267 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
D | 3x8c4-minmax-rndnu-neon-mlal-ld2r.c | 144 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_ld2r() local 234 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_ld2r() local
|
D | 3x8c4-minmax-rndnu-neon-mlal-dup.c | 144 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_dup() local 234 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_dup() local
|
D | 3x8c4-minmax-rndnu-neon-mlal-ld1r.c | 150 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_ld1r() local 243 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_ld1r() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld2r.c | 172 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local 285 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-dup.c | 172 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local 285 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local
|
D | 3x8c4s2-minmax-rndnu-neon-mull.c | 113 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4s2__neon_mull() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld1r.c | 180 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local 297 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local
|
D | 3x16c4s2-minmax-rndnu-neon-mlal.c | 195 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c4s2__neon_mlal() local 351 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c4s2__neon_mlal() local
|
D | 3x8c4-minmax-rndnu-neon-mull-dup.c | 120 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_dup() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld2r.c | 120 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_ld2r() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 133 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 3x8c4-minmax-rndnu-neon-mull-ld1r.c | 123 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c4__neon_mull_ld1r() local
|
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 3x8c4s2-minmax-rndnu-neon-mlal.c | 150 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4s2__neon_mlal() local 234 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4s2__neon_mlal() local
|
D | 4x8c4s2-minmax-rndnu-neon-mlal.c | 178 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local 284 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
D | 3x8c4-minmax-rndnu-neon-mlal-dup.c | 159 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_dup() local 249 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_dup() local
|
D | 3x8c4-minmax-rndnu-neon-mlal-ld2r.c | 159 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_ld2r() local 249 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_ld2r() local
|
D | 3x8c4-minmax-rndnu-neon-mlal-ld1r.c | 165 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_ld1r() local 258 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4__neon_mlal_ld1r() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-dup.c | 189 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local 302 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld2r.c | 189 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local 302 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local
|
D | 3x8c4s2-minmax-rndnu-neon-mull.c | 128 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c4s2__neon_mull() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld1r.c | 197 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local 314 const int16x8_t vprod2x01c1 = vmull_s8(vb01c1, va2c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local
|
D | 3x16c4s2-minmax-rndnu-neon-mlal.c | 210 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c4s2__neon_mlal() local 366 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c4s2__neon_mlal() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 150 int16x8_t vprod2x01c1 = vmull_s8(vb01c1x0, va2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|