/external/XNNPACK/src/qs8-igemm/gen/ |
D | 4x16c2-minmax-rndnu-neon-mull-dup.c | 227 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup() local 404 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup() local
|
D | 4x16c2-minmax-rndnu-neon-mull-ld4r.c | 227 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r() local 404 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r() local
|
D | 4x16c2-minmax-rndnu-neon-mlal-ld2r.c | 305 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3c2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld2r() local 513 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld2r() local 690 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld2r() local
|
D | 4x16c2-minmax-rndnu-neon-mlal-ld4r.c | 297 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3c2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld4r() local 501 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld4r() local 678 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld4r() local
|
D | 4x16c2-minmax-rndnu-neon-mlal-dup.c | 297 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3c2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_dup() local 501 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_dup() local 678 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_dup() local
|
D | 4x16c2s4-minmax-rndnu-neon-mlal.c | 286 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local 481 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local
|
D | 4x16c2-minmax-rndnu-neon-mull-ld1r.c | 239 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r() local 416 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r() local
|
D | 4x16c2-minmax-rndnu-neon-mull-ld2r.c | 231 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r() local 408 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r() local
|
D | 4x16c2-minmax-rndnu-neon-mlal-ld1r.c | 321 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3c2x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld1r() local 537 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld1r() local 714 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld1r() local
|
D | 4x16c2s4-minmax-rndnu-neon-mull.c | 219 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2s4__neon_mull() local
|
/external/XNNPACK/src/qs8-gemm/gen/ |
D | 4x16c2-minmax-rndnu-neon-mull-dup.c | 210 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup() local 387 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_dup() local
|
D | 4x16c2-minmax-rndnu-neon-mull-ld4r.c | 210 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r() local 387 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld4r() local
|
D | 4x16c2-minmax-rndnu-neon-mlal-dup.c | 279 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3c2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_dup() local 483 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_dup() local 660 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_dup() local
|
D | 4x16c2-minmax-rndnu-neon-mlal-ld2r.c | 287 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3c2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld2r() local 495 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld2r() local 672 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld2r() local
|
D | 4x16c2-minmax-rndnu-neon-mlal-ld4r.c | 279 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3c2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld4r() local 483 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld4r() local 660 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld4r() local
|
D | 4x16c2-minmax-rndnu-neon-mull-ld2r.c | 214 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r() local 391 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld2r() local
|
D | 4x16c2-minmax-rndnu-neon-mull-ld1r.c | 222 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r() local 399 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mull_ld1r() local
|
D | 4x16c2s4-minmax-rndnu-neon-mlal.c | 269 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local 464 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local
|
D | 4x16c2-minmax-rndnu-neon-mlal-ld1r.c | 303 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3c2x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld1r() local 519 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld1r() local 696 const int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2, va3c2); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2__neon_mlal_ld1r() local
|
D | 4x16c2s4-minmax-rndnu-neon-mull.c | 202 int16x8_t vprod3xCDEFc2 = vmull_s8(vbCDEFc2x0, va3x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2s4__neon_mull() local
|