| /external/XNNPACK/src/qs8-gemm/gen/ |
| D | 1x8c2-minmax-rndnu-neon-mull-dup.c | 73 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_dup() local 115 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_dup() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld4r.c | 73 const int8x8_t va0c2 = vreinterpret_s8_s16(va0.val[2]); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld4r() local 115 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld4r() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld1r.c | 76 const int8x8_t va0c2 = vreinterpret_s8_s16(va02); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld1r() local 118 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld1r() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld2r.c | 74 const int8x8_t va0c2 = vreinterpret_s8_s16(va01.val[0]); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld2r() local 116 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld2r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld2r.c | 94 const int8x8_t va0c2 = vreinterpret_s8_s16(va01.val[0]); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld2r() local 157 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld2r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld4r.c | 92 const int8x8_t va0c2 = vreinterpret_s8_s16(va0.val[2]); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld4r() local 155 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld4r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-dup.c | 92 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_dup() local 155 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_dup() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld1r.c | 98 const int8x8_t va0c2 = vreinterpret_s8_s16(va02); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld1r() local 161 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld1r() local
|
| D | 1x8c2-minmax-fp32-neon-mlal-dup.c | 132 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c2__neon_mlal_dup() local 174 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c2__neon_mlal_dup() local
|
| D | 1x8c2-minmax-fp32-neonv8-mlal-ld4r.c | 133 const int8x8_t va0c2 = vreinterpret_s8_s16(va0.val[2]); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c2__neonv8_mlal_ld4r() local 175 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_gemm_minmax_fp32_ukernel_1x8c2__neonv8_mlal_ld4r() local
|
| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 1x8c2-minmax-rndnu-neon-mull-ld2r.c | 85 const int8x8_t va0c2 = vreinterpret_s8_s16(va01.val[0]); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld2r() local 127 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld2r() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-dup.c | 84 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_dup() local 126 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_dup() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld4r.c | 84 const int8x8_t va0c2 = vreinterpret_s8_s16(va0.val[2]); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld4r() local 126 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld4r() local
|
| D | 1x8c2-minmax-rndnu-neon-mull-ld1r.c | 87 const int8x8_t va0c2 = vreinterpret_s8_s16(va02); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld1r() local 129 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mull_ld1r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-dup.c | 105 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_dup() local 168 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_dup() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld4r.c | 105 const int8x8_t va0c2 = vreinterpret_s8_s16(va0.val[2]); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld4r() local 168 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld4r() local
|
| D | 2x8c2-minmax-rndnu-neon-mull-ld1r.c | 111 const int8x8_t va0c2 = vreinterpret_s8_s16(va02); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld1r() local 174 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x8c2__neon_mull_ld1r() local
|
| D | 1x16c2-minmax-rndnu-neon-mull-dup.c | 102 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16c2__neon_mull_dup() local 166 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16c2__neon_mull_dup() local
|
| D | 1x8c2-minmax-fp32-neonv8-mlal-ld4r.c | 145 const int8x8_t va0c2 = vreinterpret_s8_s16(va0.val[2]); in xnn_qs8_igemm_minmax_fp32_ukernel_1x8c2__neonv8_mlal_ld4r() local 187 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_fp32_ukernel_1x8c2__neonv8_mlal_ld4r() local
|
| D | 1x8c2-minmax-rndnu-neon-mlal-ld4r.c | 144 const int8x8_t va0c2 = vreinterpret_s8_s16(va0.val[2]); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mlal_ld4r() local 186 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x8c2__neon_mlal_ld4r() local
|
| /external/XNNPACK/src/f16-gemm/gen/ |
| D | 1x8-minmax-neonfp16arith-ld64.c | 74 const float16x8_t va0c2 = vdupq_lane_f16(va0, 2); in xnn_f16_gemm_minmax_ukernel_1x8__neonfp16arith_ld64() local
|
| /external/XNNPACK/src/f16-gemm/gen-inc/ |
| D | 1x8inc-minmax-neonfp16arith-ld64.c | 76 const float16x8_t va0c2 = vdupq_lane_f16(va0, 2); in xnn_f16_gemminc_minmax_ukernel_1x8__neonfp16arith_ld64() local
|
| /external/XNNPACK/src/qc8-igemm/gen/ |
| D | 1x8c2-minmax-fp32-neonv8-mlal-ld4r.c | 145 const int8x8_t va0c2 = vreinterpret_s8_s16(va0.val[2]); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c2__neonv8_mlal_ld4r() local 187 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c2__neonv8_mlal_ld4r() local
|
| D | 1x8c2-minmax-fp32-neon-mlal-dup.c | 144 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qc8_igemm_minmax_fp32_ukernel_1x8c2__neon_mlal_dup() local 186 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); 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 | 132 const int8x8_t va0c2 = vreinterpret_s8_s16(va0.val[2]); in xnn_qc8_gemm_minmax_fp32_ukernel_1x8c2__neon_mlal_ld4r() local 174 const int8x8_t va0c2 = vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)); in xnn_qc8_gemm_minmax_fp32_ukernel_1x8c2__neon_mlal_ld4r() local
|