| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 4x8c8-minmax-neon-mlal-padal.c | 137 int16x8_t vprod3x0 = vmull_s8(vb0x0, va3x0); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
141 vprod3x0 = vmlal_s8(vprod3x0, vb0x1, va3x1); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal()
145 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal()
252 const int16x8_t vprod3x0 = vmull_s8(vb0, va3); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
256 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal()
|
| D | 4x8c16-minmax-neon-mlal-padal.c | 132 int16x8_t vprod3x0 = vmull_s8(vget_low_s8(vb0), vget_low_s8(va3)); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal() local
136 vprod3x0 = vmlal_s8(vprod3x0, vget_high_s8(vb0), vget_high_s8(va3)); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
140 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_igemm_minmax_ukernel_4x8c16__neon_mlal_padal()
|
| D | 4x16c8-minmax-neon-mlal-padal.c | 177 int16x8_t vprod3x0 = vmull_s8(vb0x0, va3x0); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
181 vprod3x0 = vmlal_s8(vprod3x0, vb0x1, va3x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
185 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
396 const int16x8_t vprod3x0 = vmull_s8(vb0, va3); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
400 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
|
| D | 4x16c16-minmax-neon-mlal-padal.c | 172 int16x8_t vprod3x0 = vmull_s8(vget_low_s8(vb0), vget_low_s8(va3)); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal() local
176 vprod3x0 = vmlal_s8(vprod3x0, vget_high_s8(vb0), vget_high_s8(va3)); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
180 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_igemm_minmax_ukernel_4x16c16__neon_mlal_padal()
|
| D | 4x8c8-minmax-neon-mull-padal.c | 125 const int16x8_t vprod3x0 = vmull_s8(vb0, va3); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal() local
129 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mull_padal()
|
| D | 4x16c8-minmax-neon-mull-padal.c | 157 const int16x8_t vprod3x0 = vmull_s8(vb0, va3); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal() local
161 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mull_padal()
|
| /external/XNNPACK/src/qs8-gemm/gen/ |
| D | 4x8c8-minmax-neon-mlal-padal.c | 120 int16x8_t vprod3x0 = vmull_s8(vb0x0, va3x0); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
124 vprod3x0 = vmlal_s8(vprod3x0, vb0x1, va3x1); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal()
128 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal()
235 const int16x8_t vprod3x0 = vmull_s8(vb0, va3); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
239 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal()
|
| D | 4x8c16-minmax-neon-mlal-padal.c | 115 int16x8_t vprod3x0 = vmull_s8(vget_low_s8(vb0), vget_low_s8(va3)); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal() local
119 vprod3x0 = vmlal_s8(vprod3x0, vget_high_s8(vb0), vget_high_s8(va3)); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
123 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_gemm_minmax_ukernel_4x8c16__neon_mlal_padal()
|
| D | 4x16c8-minmax-neon-mlal-padal.c | 160 int16x8_t vprod3x0 = vmull_s8(vb0x0, va3x0); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
164 vprod3x0 = vmlal_s8(vprod3x0, vb0x1, va3x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
168 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
379 const int16x8_t vprod3x0 = vmull_s8(vb0, va3); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() local
383 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
|
| D | 4x16c16-minmax-neon-mlal-padal.c | 155 int16x8_t vprod3x0 = vmull_s8(vget_low_s8(vb0), vget_low_s8(va3)); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal() local
159 vprod3x0 = vmlal_s8(vprod3x0, vget_high_s8(vb0), vget_high_s8(va3)); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
163 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_gemm_minmax_ukernel_4x16c16__neon_mlal_padal()
|
| D | 4x8c8-minmax-neon-mull-padal.c | 108 const int16x8_t vprod3x0 = vmull_s8(vb0, va3); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal() local
112 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mull_padal()
|
| D | 4x16c8-minmax-neon-mull-padal.c | 140 const int16x8_t vprod3x0 = vmull_s8(vb0, va3); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal() local
144 vacc3x0 = vpadalq_s16(vacc3x0, vprod3x0); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mull_padal()
|