| /external/XNNPACK/src/qs8-vaddc/ |
| D | neon-ld64.c.in | 116 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout${ABC[0:8]}), 0); outpu…
130 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout${ABC[0:8]}), 0); outpu…
|
| /external/XNNPACK/src/qs8-vadd/ |
| D | neon-ld64.c.in | 123 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout${ABC[0:8]}), 0); outpu…
137 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout${ABC[0:8]}), 0); outpu…
|
| /external/XNNPACK/src/qs8-vaddc/gen/ |
| D | minmax-neon-ld64-x8.c | 89 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout01234567), 0); output +… in xnn_qs8_vaddc_minmax_ukernel__neon_ld64_x8()
|
| D | minmax-neon-ld64-x16.c | 102 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout01234567), 0); output +… in xnn_qs8_vaddc_minmax_ukernel__neon_ld64_x16()
|
| D | minmax-neon-ld64-x24.c | 115 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout01234567), 0); output +… in xnn_qs8_vaddc_minmax_ukernel__neon_ld64_x24()
|
| D | minmax-neon-ld64-x32.c | 124 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout01234567), 0); output +… in xnn_qs8_vaddc_minmax_ukernel__neon_ld64_x32()
|
| /external/XNNPACK/src/qs8-gemm/gen/ |
| D | 1x8c4-minmax-neondot.c | 131 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_gemm_minmax_ukernel_1x8c4__neondot()
|
| D | 1x16c4-minmax-neondot.c | 158 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot()
|
| D | 1x8c8-minmax-neon-mull-padal.c | 153 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_gemm_minmax_ukernel_1x8c8__neon_mull_padal()
|
| D | 1x8c2-minmax-neon-mull-padal-dup.c | 156 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_gemm_minmax_ukernel_1x8c2__neon_mull_padal_dup()
|
| D | 1x8-minmax-neon-mlal-lane.c | 197 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_gemm_minmax_ukernel_1x8__neon_mlal_lane()
|
| D | 1x8c16-minmax-neon-mlal-padal.c | 161 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_gemm_minmax_ukernel_1x8c16__neon_mlal_padal()
|
| D | 1x8-minmax-neon-mull-addw-dup.c | 193 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_gemm_minmax_ukernel_1x8__neon_mull_addw_dup()
|
| /external/XNNPACK/src/qs8-vadd/gen/ |
| D | minmax-neon-ld64-x8.c | 94 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout01234567), 0); output +… in xnn_qs8_vadd_minmax_ukernel__neon_ld64_x8()
|
| D | minmax-neon-ld64-x16.c | 111 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout01234567), 0); output +… in xnn_qs8_vadd_minmax_ukernel__neon_ld64_x16()
|
| D | minmax-neon-ld64-x24.c | 128 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout01234567), 0); output +… in xnn_qs8_vadd_minmax_ukernel__neon_ld64_x24()
|
| D | minmax-neon-ld64-x32.c | 141 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout01234567), 0); output +… in xnn_qs8_vadd_minmax_ukernel__neon_ld64_x32()
|
| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 1x8c4-minmax-neondot.c | 138 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_igemm_minmax_ukernel_1x8c4__neondot()
|
| D | 1x16c4-minmax-neondot.c | 165 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot()
|
| D | 1x8c8-minmax-neon-mull-padal.c | 167 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_igemm_minmax_ukernel_1x8c8__neon_mull_padal()
|
| D | 1x8c2-minmax-neon-mull-padal-dup.c | 170 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_igemm_minmax_ukernel_1x8c2__neon_mull_padal_dup()
|
| D | 1x8-minmax-neon-mlal-lane.c | 210 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_igemm_minmax_ukernel_1x8__neon_mlal_lane()
|
| D | 1x8-minmax-neon-mull-addw-dup.c | 207 … vst1_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpret_u16_s8(vout0x01234567), 0); c0 += 2; in xnn_qs8_igemm_minmax_ukernel_1x8__neon_mull_addw_dup()
|
| /external/XNNPACK/src/qs8-dwconv/ |
| D | unipass-neon-mul16.c.in | 182 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout${ABC[0:8]}), 0); outpu…
196 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout${ABC[0:8]}), 0); outpu…
|
| /external/XNNPACK/src/qs8-gavgpool/ |
| D | unipass-neon.c.in | 245 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout${ABC[0:8]}), 0); outpu…
259 …vst1_lane_u16(__builtin_assume_aligned(output, 1), vreinterpret_u16_s8(vout${ABC[0:8]}), 0); outpu…
|