/external/XNNPACK/src/qs8-gemm/gen/ |
D | 2x16c2s4-minmax-rndnu-neon-mlal.c | 201 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2s4__neon_mlal() local 298 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2s4__neon_mlal() local
|
D | 3x16c2s4-minmax-rndnu-neon-mlal.c | 264 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2s4__neon_mlal() local 398 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2s4__neon_mlal() local
|
D | 2x16c2s4-minmax-rndnu-neon-mull.c | 148 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2s4__neon_mull() local
|
D | 2x16c2-minmax-rndnu-neon-mlal-dup.c | 208 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_dup() local 312 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_dup() local
|
D | 2x16c2-minmax-rndnu-neon-mlal-ld4r.c | 208 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld4r() local 312 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld4r() local
|
D | 2x16c2-minmax-rndnu-neon-mlal-ld2r.c | 212 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld2r() local 318 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld2r() local
|
D | 2x16c2-minmax-rndnu-neon-mlal-ld1r.c | 220 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld1r() local 330 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld1r() local
|
D | 4x16c2s4-minmax-rndnu-neon-mlal.c | 327 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local 498 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local
|
D | 3x16c2s4-minmax-rndnu-neon-mull.c | 192 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2s4__neon_mull() local
|
D | 3x16c2-minmax-rndnu-neon-mlal-dup.c | 273 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_dup() local 415 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_dup() local
|
D | 3x16c2-minmax-rndnu-neon-mlal-ld4r.c | 273 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld4r() local 415 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld4r() local
|
D | 3x16c2-minmax-rndnu-neon-mlal-ld2r.c | 279 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld2r() local 424 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld2r() local
|
D | 3x16c2-minmax-rndnu-neon-mlal-ld1r.c | 291 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld1r() local 442 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld1r() local
|
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 2x16c2s4-minmax-rndnu-neon-mlal.c | 214 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2s4__neon_mlal() local 311 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2s4__neon_mlal() local
|
D | 3x16c2s4-minmax-rndnu-neon-mlal.c | 279 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2s4__neon_mlal() local 413 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2s4__neon_mlal() local
|
D | 2x16c2s4-minmax-rndnu-neon-mull.c | 161 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2s4__neon_mull() local
|
D | 2x16c2-minmax-rndnu-neon-mlal-ld4r.c | 222 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld4r() local 326 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld4r() local
|
D | 2x16c2-minmax-rndnu-neon-mlal-dup.c | 222 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_dup() local 326 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_dup() local
|
D | 2x16c2-minmax-rndnu-neon-mlal-ld2r.c | 226 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld2r() local 332 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld2r() local
|
D | 2x16c2-minmax-rndnu-neon-mlal-ld1r.c | 234 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld1r() local 344 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c2__neon_mlal_ld1r() local
|
D | 4x16c2s4-minmax-rndnu-neon-mlal.c | 344 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local 515 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local
|
D | 3x16c2s4-minmax-rndnu-neon-mull.c | 207 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2s4__neon_mull() local
|
D | 3x16c2-minmax-rndnu-neon-mlal-ld4r.c | 289 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld4r() local 431 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld4r() local
|
D | 3x16c2-minmax-rndnu-neon-mlal-dup.c | 289 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_dup() local 431 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_dup() local
|
D | 3x16c2-minmax-rndnu-neon-mlal-ld1r.c | 307 int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3x0, va1c3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld1r() local 458 const int16x8_t vprod1xCDEFc3 = vmull_s8(vbCDEFc3, va1c3); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c2__neon_mlal_ld1r() local
|