/external/XNNPACK/src/qs8-gemm/gen/ |
D | 4x8c4s2-minmax-rndnu-neon-mlal.c | 162 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local 268 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld2r.c | 173 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local 286 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-dup.c | 173 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local 286 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld1r.c | 181 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local 298 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 134 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 4x16c4s2-minmax-rndnu-neon-mlal.c | 238 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local 436 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local
|
D | 4x8c4-minmax-rndnu-neon-mull-ld2r.c | 142 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld2r() local
|
D | 4x8c4-minmax-rndnu-neon-mull-dup.c | 142 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mull_dup() local
|
D | 4x8c4-minmax-rndnu-neon-mull-ld1r.c | 146 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld1r() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-dup.c | 249 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_dup() local 454 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_dup() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-ld1r.c | 257 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld1r() local 466 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld1r() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-ld2r.c | 249 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld2r() local 454 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, 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 | 179 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local 285 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-dup.c | 190 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local 303 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld2r.c | 190 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local 303 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local
|
D | 4x8c4-minmax-rndnu-neon-mlal-ld1r.c | 198 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local 315 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld1r() local
|
D | 4x8c4s2-minmax-rndnu-neon-mull.c | 151 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
D | 4x16c4s2-minmax-rndnu-neon-mlal.c | 255 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local 453 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local
|
D | 4x8c4-minmax-rndnu-neon-mull-ld2r.c | 159 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld2r() local
|
D | 4x8c4-minmax-rndnu-neon-mull-dup.c | 159 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mull_dup() local
|
D | 4x8c4-minmax-rndnu-neon-mull-ld1r.c | 163 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mull_ld1r() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-ld1r.c | 274 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld1r() local 483 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld1r() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-dup.c | 266 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_dup() local 471 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_dup() local
|
D | 4x16c4-minmax-rndnu-neon-mlal-ld2r.c | 266 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3c1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld2r() local 471 const int16x8_t vprod3x01c1 = vmull_s8(vb01c1, va3c1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4__neon_mlal_ld2r() local
|
D | 4x16c4s2-minmax-rndnu-neon-mull.c | 207 int16x8_t vprod3x01c1 = vmull_s8(vb01c1x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4s2__neon_mull() local
|