| /external/XNNPACK/src/qs8-gemm/gen/ |
| D | 1x8c16-minmax-rndnu-neon-mlal.c | 96 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c16__neon_mlal() local
107 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c16__neon_mlal() local
|
| D | 1x8c8-minmax-rndnu-neon-mull.c | 88 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c8__neon_mull() local
100 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c8__neon_mull() local
|
| D | 1x8c8-minmax-fp32-neonv8-mlal.c | 138 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c8__neonv8_mlal() local
150 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c8__neonv8_mlal() local
|
| D | 1x8c8-minmax-rndnu-neon-mlal.c | 137 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c8__neon_mlal() local
149 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c8__neon_mlal() local
|
| D | 1x8c8-minmax-fp32-neon-mlal.c | 137 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c8__neon_mlal() local
149 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c8__neon_mlal() local
|
| D | 2x8c8-minmax-rndnu-neon-mull.c | 119 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c8__neon_mull() local
137 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c8__neon_mull() local
|
| D | 2x8c16-minmax-rndnu-neon-mlal.c | 135 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c16__neon_mlal() local
152 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c16__neon_mlal() local
|
| D | 1x16c8-minmax-rndnu-neon-mull.c | 120 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c8__neon_mull() local
138 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c8__neon_mull() local
|
| D | 1x16c16-minmax-rndnu-neon-mlal.c | 136 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c16__neon_mlal() local
153 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c16__neon_mlal() local
|
| D | 1x8c4s2-minmax-rndnu-neon-mull.c | 87 const int32x2_t vsum0x01 = vpadd_s32(vget_low_s32(vacc0x01), vget_high_s32(vacc0x01)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c4s2__neon_mull() local
|
| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 1x8c8-minmax-rndnu-neon-mull.c | 102 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c8__neon_mull() local
114 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c8__neon_mull() local
|
| D | 1x8c16-minmax-rndnu-neon-mlal.c | 110 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c16__neon_mlal() local
121 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c16__neon_mlal() local
|
| D | 1x8c8-minmax-fp32-neonv8-mlal.c | 152 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_fp32_ukernel_1x8c8__neonv8_mlal() local
164 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_fp32_ukernel_1x8c8__neonv8_mlal() local
|
| D | 1x8c8-minmax-rndnu-neon-mlal.c | 151 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c8__neon_mlal() local
163 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c8__neon_mlal() local
|
| D | 1x8c8-minmax-fp32-neon-mlal.c | 151 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_fp32_ukernel_1x8c8__neon_mlal() local
163 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_fp32_ukernel_1x8c8__neon_mlal() local
|
| D | 1x16c8-minmax-rndnu-neon-mull.c | 134 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16c8__neon_mull() local
152 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16c8__neon_mull() local
|
| D | 2x8c8-minmax-rndnu-neon-mull.c | 135 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c8__neon_mull() local
153 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c8__neon_mull() local
|
| D | 2x8c16-minmax-rndnu-neon-mlal.c | 151 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c16__neon_mlal() local
168 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c16__neon_mlal() local
|
| D | 1x16c16-minmax-rndnu-neon-mlal.c | 150 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16c16__neon_mlal() local
167 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16c16__neon_mlal() local
|
| D | 2x8c8-minmax-rndnu-neon-mlal.c | 210 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c8__neon_mlal() local
228 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c8__neon_mlal() local
|
| /external/XNNPACK/src/qc8-gemm/gen/ |
| D | 1x8c8-minmax-fp32-neon-mlal.c | 137 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qc8_gemm_minmax_fp32_ukernel_1x8c8__neon_mlal() local
149 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qc8_gemm_minmax_fp32_ukernel_1x8c8__neon_mlal() local
|
| D | 1x8c8-minmax-fp32-neonv8-mlal.c | 138 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qc8_gemm_minmax_fp32_ukernel_1x8c8__neonv8_mlal() local
150 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qc8_gemm_minmax_fp32_ukernel_1x8c8__neonv8_mlal() local
|
| /external/XNNPACK/src/qc8-igemm/gen/ |
| D | 1x8c8-minmax-fp32-neon-mlal.c | 151 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c8__neon_mlal() local
163 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c8__neon_mlal() local
|
| D | 1x8c8-minmax-fp32-neonv8-mlal.c | 152 const int32x4_t vsum0x01 = vpaddq_s32(vacc0x0, vacc0x1); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c8__neonv8_mlal() local
164 const int32x2_t vsum0x01 = vpadd_s32(vpsum0x0, vpsum0x1); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c8__neonv8_mlal() local
|
| /external/pytorch/aten/src/ATen/native/quantized/cpu/qnnpack/src/q8gemm/ |
| D | 4x-sumrows-neon.c | 100 const uint32x2_t vsum0x01 = in pytorch_q8sumrows_ukernel_4x__neon() local
|