| /external/XNNPACK/src/qs8-gemm/gen/ |
| D | 1x8c2-minmax-rndnu-neon-mull-dup.c | 67 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_dup() local
105 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_dup() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld4r.c | 67 const int8x8_t va0c1 = vreinterpret_s8_s16(va0.val[1]); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld4r() local
105 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld4r() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld1r.c | 70 const int8x8_t va0c1 = vreinterpret_s8_s16(va01); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld1r() local
108 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld1r() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld2r.c | 68 const int8x8_t va0c1 = vreinterpret_s8_s16(va00.val[1]); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld2r() local
106 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld2r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld2r.c | 83 const int8x8_t va0c1 = vreinterpret_s8_s16(va00.val[1]); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld2r() local
142 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld2r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld4r.c | 81 const int8x8_t va0c1 = vreinterpret_s8_s16(va0.val[1]); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld4r() local
140 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld4r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-dup.c | 81 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_dup() local
140 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_dup() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld1r.c | 87 const int8x8_t va0c1 = vreinterpret_s8_s16(va01); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld1r() local
146 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld1r() local
|
| D | 1x8c2-minmax-fp32-neon-mlal-dup.c | 126 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c2__neon_mlal_dup() local
164 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c2__neon_mlal_dup() local
|
| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 1x8c2-minmax-rndnu-neon-mull-ld2r.c | 79 const int8x8_t va0c1 = vreinterpret_s8_s16(va00.val[1]); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld2r() local
117 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld2r() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-dup.c | 78 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_dup() local
116 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_dup() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld4r.c | 78 const int8x8_t va0c1 = vreinterpret_s8_s16(va0.val[1]); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld4r() local
116 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld4r() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld1r.c | 81 const int8x8_t va0c1 = vreinterpret_s8_s16(va01); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld1r() local
119 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld1r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-dup.c | 94 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_dup() local
153 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_dup() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld4r.c | 94 const int8x8_t va0c1 = vreinterpret_s8_s16(va0.val[1]); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld4r() local
153 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld4r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld1r.c | 100 const int8x8_t va0c1 = vreinterpret_s8_s16(va01); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld1r() local
159 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld1r() local
|
| /external/XNNPACK/src/f16-gemm/gen/ |
| D | 1x8-minmax-neonfp16arith-ld64.c | 65 const float16x8_t va0c1 = vdupq_lane_f16(va0, 1); in xnn_f16_gemm_minmax_ukernel_1x8__neonfp16arith_ld64() local
|
| /external/XNNPACK/src/f32-gemm/gen/ |
| D | 1x8-minmax-neon-dup-ld64.c | 59 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemm_minmax_ukernel_1x8__neon_dup_ld64() local
|
| D | 1x8-minmax-neonfma-dup-ld64.c | 59 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemm_minmax_ukernel_1x8__neonfma_dup_ld64() local
|
| /external/XNNPACK/src/f32-gemm/gen-inc/ |
| D | 1x8inc-minmax-neonfma-dup-ld64.c | 61 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemminc_minmax_ukernel_1x8__neonfma_dup_ld64() local
|
| D | 1x8inc-minmax-neon-dup-ld64.c | 61 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemminc_minmax_ukernel_1x8__neon_dup_ld64() local
|
| /external/XNNPACK/src/f16-gemm/gen-inc/ |
| D | 1x8inc-minmax-neonfp16arith-ld64.c | 67 const float16x8_t va0c1 = vdupq_lane_f16(va0, 1); in xnn_f16_gemminc_minmax_ukernel_1x8__neonfp16arith_ld64() local
|
| /external/XNNPACK/src/qc8-igemm/gen/ |
| D | 1x8c2-minmax-fp32-neonv8-mlal-ld4r.c | 139 const int8x8_t va0c1 = vreinterpret_s8_s16(va0.val[1]); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c2__neonv8_mlal_ld4r() local
177 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c2__neonv8_mlal_ld4r() local
|
| D | 1x8c2-minmax-fp32-neon-mlal-dup.c | 138 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c2__neon_mlal_dup() local
176 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c2__neon_mlal_dup() local
|
| /external/XNNPACK/src/qc8-gemm/gen/ |
| D | 1x8c2-minmax-fp32-neon-mlal-ld4r.c | 126 const int8x8_t va0c1 = vreinterpret_s8_s16(va0.val[1]); in xnn_qc8_gemm_minmax_fp32_ukernel_1x8c2__neon_mlal_ld4r() local
164 const int8x8_t va0c1 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)); in xnn_qc8_gemm_minmax_fp32_ukernel_1x8c2__neon_mlal_ld4r() local
|