/external/XNNPACK/src/qs8-gemm/gen/ |
D | 4x8c4s2-minmax-rndnu-neon-mlal.c | 188 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local 284 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld2r.c | 199 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local 302 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-dup.c | 199 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local 302 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld1r.c | 207 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local 314 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 150 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 4x16c4s2-minmax-rndnu-neon-mlal.c | 264 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local 452 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local
|
D | 4x8c4-minmax-rndnu-neon-mull-ld2r.c | 158 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld2r() local
|
D | 4x8c4-minmax-rndnu-neon-mull-dup.c | 158 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mull_dup() local
|
D | 4x8c4-minmax-rndnu-neon-mull-ld1r.c | 162 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld1r() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-dup.c | 275 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_dup() local 470 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_dup() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-ld1r.c | 283 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld1r() local 482 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld1r() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-ld2r.c | 275 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld2r() local 470 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld2r() local
|
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 4x8c4s2-minmax-rndnu-neon-mlal.c | 205 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local 301 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-dup.c | 216 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local 319 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld2r.c | 216 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local 319 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld1r.c | 224 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local 331 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 167 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 4x16c4s2-minmax-rndnu-neon-mlal.c | 281 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local 469 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local
|
D | 4x8c4-minmax-rndnu-neon-mull-ld2r.c | 175 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld2r() local
|
D | 4x8c4-minmax-rndnu-neon-mull-dup.c | 175 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mull_dup() local
|
D | 4x8c4-minmax-rndnu-neon-mull-ld1r.c | 179 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld1r() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-ld1r.c | 300 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld1r() local 499 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld1r() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-dup.c | 292 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_dup() local 487 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_dup() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-ld2r.c | 292 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld2r() local 487 const int16x8_t vprod3x45c1 = vmull_s8(vb45c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld2r() local
|
D | 4x16c4s2-minmax-rndnu-neon-mull.c | 223 int16x8_t vprod3x45c1 = vmull_s8(vb45c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4s2__neon_mull() local
|