/external/XNNPACK/src/qs8-igemm/gen/ |
D | 3x16c8-minmax-neon-mull-padal.c | 269 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal() local 282 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal() 360 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal() local 361 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mull_padal()
|
D | 3x16c16-minmax-neon-mlal-padal.c | 317 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal() local 330 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal() 408 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal() local 409 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_igemm_minmax_ukernel_3x16c16__neon_mlal_padal()
|
D | 4x16c8-minmax-neon-mull-padal.c | 326 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal() local 347 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal() 429 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal() local 430 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
|
D | 4x16c16-minmax-neon-mlal-padal.c | 390 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal() local 411 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal() 493 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal() local 494 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
|
D | 3x16c8-minmax-neon-mlal-padal.c | 458 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() local 471 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() 549 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() local 550 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal()
|
D | 4x16c8-minmax-neon-mlal-padal.c | 565 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() local 586 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() 668 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() local 669 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
|
/external/XNNPACK/src/qs8-gemm/gen/ |
D | 3x16c8-minmax-neon-mull-padal.c | 251 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal() local 264 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal() 342 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal() local 343 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mull_padal()
|
D | 3x16c16-minmax-neon-mlal-padal.c | 299 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal() local 312 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal() 390 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal() local 391 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_gemm_minmax_ukernel_3x16c16__neon_mlal_padal()
|
D | 4x16c8-minmax-neon-mull-padal.c | 306 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal() local 327 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal() 409 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal() local 410 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
|
D | 3x16c8-minmax-neon-mlal-padal.c | 440 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() local 453 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() 531 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() local 532 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
|
D | 4x16c16-minmax-neon-mlal-padal.c | 370 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal() local 391 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal() 473 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal() local 474 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
|
D | 4x16c8-minmax-neon-mlal-padal.c | 545 const int32x4_t vsum2xAB = vpaddq_s32(vacc2x10, vacc2x11); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() local 566 int32x4_t vacc2x89AB = vpaddq_s32(vsum2x89, vsum2xAB); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() 648 const int32x2_t vsum2xAB = vpadd_s32(vpsum2xA, vpsum2xB); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() local 649 int32x4_t vacc2x89AB = vcombine_s32(vsum2x89, vsum2xAB ); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
|