| /external/XNNPACK/src/qs8-gemm/gen/ |
| D | 2x8c8-minmax-fp32-neonv8-mlal.c | 115 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_gemm_minmax_fp32_ukernel_2x8c8__neonv8_mlal() local
172 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_gemm_minmax_fp32_ukernel_2x8c8__neonv8_mlal() local
|
| D | 2x8c8-minmax-fp32-neon-mlal.c | 114 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_gemm_minmax_fp32_ukernel_2x8c8__neon_mlal() local
171 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_gemm_minmax_fp32_ukernel_2x8c8__neon_mlal() local
|
| D | 2x8c8-minmax-rndnu-neon-mlal.c | 114 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c8__neon_mlal() local
171 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c8__neon_mlal() local
|
| D | 3x8c8-minmax-rndnu-neon-mlal.c | 142 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
220 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
|
| D | 2x8c16-minmax-rndnu-neon-mlal.c | 107 int16x8_t vprod1x4 = vmull_s8(vget_low_s8(vb4), vget_low_s8(va1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c16__neon_mlal() local
|
| D | 2x8c8-minmax-rndnu-neon-mull.c | 96 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c8__neon_mull() local
|
| D | 4x8c8-minmax-rndnu-neon-mlal.c | 170 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
269 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
| D | 2x16c8-minmax-rndnu-neon-mlal.c | 138 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() local
251 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() local
|
| D | 3x8c8-minmax-rndnu-neon-mull.c | 119 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local
|
| D | 3x8c16-minmax-rndnu-neon-mlal.c | 134 int16x8_t vprod1x4 = vmull_s8(vget_low_s8(vb4), vget_low_s8(va1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local
|
| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 2x8c8-minmax-rndnu-neon-mlal.c | 127 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c8__neon_mlal() local
184 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c8__neon_mlal() local
|
| D | 2x8c8-minmax-fp32-neonv8-mlal.c | 128 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_igemm_minmax_fp32_ukernel_2x8c8__neonv8_mlal() local
185 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_igemm_minmax_fp32_ukernel_2x8c8__neonv8_mlal() local
|
| D | 2x8c8-minmax-fp32-neon-mlal.c | 127 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_igemm_minmax_fp32_ukernel_2x8c8__neon_mlal() local
184 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_igemm_minmax_fp32_ukernel_2x8c8__neon_mlal() local
|
| D | 3x8c8-minmax-rndnu-neon-mlal.c | 157 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
235 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mlal() local
|
| D | 4x8c8-minmax-rndnu-neon-mlal.c | 187 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
286 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c8__neon_mlal() local
|
| D | 2x8c8-minmax-rndnu-neon-mull.c | 109 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c8__neon_mull() local
|
| D | 2x8c16-minmax-rndnu-neon-mlal.c | 120 int16x8_t vprod1x4 = vmull_s8(vget_low_s8(vb4), vget_low_s8(va1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c16__neon_mlal() local
|
| D | 2x16c8-minmax-rndnu-neon-mlal.c | 151 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() local
264 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() local
|
| D | 3x8c16-minmax-rndnu-neon-mlal.c | 149 int16x8_t vprod1x4 = vmull_s8(vget_low_s8(vb4), vget_low_s8(va1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c16__neon_mlal() local
|
| D | 3x8c8-minmax-rndnu-neon-mull.c | 134 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x8c8__neon_mull() local
|
| D | 3x16c8-minmax-rndnu-neon-mlal.c | 189 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() local
347 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() local
|
| /external/XNNPACK/src/qc8-gemm/gen/ |
| D | 2x8c8-minmax-fp32-neon-mlal.c | 114 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qc8_gemm_minmax_fp32_ukernel_2x8c8__neon_mlal() local
171 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qc8_gemm_minmax_fp32_ukernel_2x8c8__neon_mlal() local
|
| D | 2x8c8-minmax-fp32-neonv8-mlal.c | 115 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qc8_gemm_minmax_fp32_ukernel_2x8c8__neonv8_mlal() local
172 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qc8_gemm_minmax_fp32_ukernel_2x8c8__neonv8_mlal() local
|
| /external/XNNPACK/src/qc8-igemm/gen/ |
| D | 2x8c8-minmax-fp32-neonv8-mlal.c | 128 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qc8_igemm_minmax_fp32_ukernel_2x8c8__neonv8_mlal() local
185 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qc8_igemm_minmax_fp32_ukernel_2x8c8__neonv8_mlal() local
|
| D | 2x8c8-minmax-fp32-neon-mlal.c | 127 int16x8_t vprod1x4 = vmull_s8(vb4x0, va1x0); in xnn_qc8_igemm_minmax_fp32_ukernel_2x8c8__neon_mlal() local
184 const int16x8_t vprod1x4 = vmull_s8(vb4, va1); in xnn_qc8_igemm_minmax_fp32_ukernel_2x8c8__neon_mlal() local
|