| /external/XNNPACK/src/qs8-gemm/gen/ |
| D | 4x8c2s4-minmax-rndnu-neon-mlal.c | 79 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c2s4__neon_mlal() local
226 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c2s4__neon_mlal() local
|
| D | 4x8c4s2-minmax-rndnu-neon-mlal.c | 87 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
218 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
| D | 4x8c2s4-minmax-rndnu-neon-mull.c | 76 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c2s4__neon_mull() local
|
| D | 4x16c2s4-minmax-rndnu-neon-mlal.c | 87 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local
346 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local
|
| D | 4x8c4s2-minmax-rndnu-neon-mull.c | 84 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
| D | 4x16c4s2-minmax-rndnu-neon-mlal.c | 103 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local
346 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local
|
| D | 4x16c2s4-minmax-rndnu-neon-mull.c | 84 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c2s4__neon_mull() local
|
| D | 4x8c4-minmax-rndnu-neon-mlal-dup.c | 88 const int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_gemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_dup() local
|
| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 4x8c2s4-minmax-rndnu-neon-mlal.c | 96 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c2s4__neon_mlal() local
243 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c2s4__neon_mlal() local
|
| D | 4x8c4s2-minmax-rndnu-neon-mlal.c | 104 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
235 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mlal() local
|
| D | 4x8c2s4-minmax-rndnu-neon-mull.c | 93 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c2s4__neon_mull() local
|
| D | 4x16c2s4-minmax-rndnu-neon-mlal.c | 104 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local
363 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2s4__neon_mlal() local
|
| D | 4x8c4s2-minmax-rndnu-neon-mull.c | 101 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4s2__neon_mull() local
|
| D | 4x16c4s2-minmax-rndnu-neon-mlal.c | 120 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local
363 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c4s2__neon_mlal() local
|
| D | 4x16c2s4-minmax-rndnu-neon-mull.c | 101 int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c2s4__neon_mull() local
|
| D | 4x8c2-minmax-rndnu-neon-mlal-dup.c | 96 const int8x8_t va3x0 = vld1_s8(a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c2__neon_mlal_dup() local
|
| D | 4x8c4-minmax-rndnu-neon-mlal-ld2r.c | 105 const int32x2x2_t va3x0 = vld2_dup_s32((const void*)a3); a3 += 8; in xnn_qs8_igemm_minmax_rndnu_ukernel_4x8c4__neon_mlal_ld2r() local
|
| /external/XNNPACK/src/bf16-gemm/gen/ |
| D | 4x4c8-minmax-neonbf16-bfdot.c | 130 … const bfloat16x8_t va3x0 = vreinterpretq_bf16_u16(vbicq_u16(vreinterpretq_u16_bf16(va3), vm0)); in xnn_bf16_gemm_minmax_ukernel_4x4c8__neonbf16_bfdot() local
|
| D | 4x4c8-minmax-neonbf16-bfmlal.c | 150 … const bfloat16x8_t va3x0 = vreinterpretq_bf16_u16(vbicq_u16(vreinterpretq_u16_bf16(va3), vm0)); in xnn_bf16_gemm_minmax_ukernel_4x4c8__neonbf16_bfmlal() local
|
| D | 5x4c8-minmax-neonbf16-bfmlal.c | 170 … const bfloat16x8_t va3x0 = vreinterpretq_bf16_u16(vbicq_u16(vreinterpretq_u16_bf16(va3), vm0)); in xnn_bf16_gemm_minmax_ukernel_5x4c8__neonbf16_bfmlal() local
|
| D | 5x4c8-minmax-neonbf16-bfdot.c | 146 … const bfloat16x8_t va3x0 = vreinterpretq_bf16_u16(vbicq_u16(vreinterpretq_u16_bf16(va3), vm0)); in xnn_bf16_gemm_minmax_ukernel_5x4c8__neonbf16_bfdot() local
|
| D | 4x4c8-minmax-neonfma-shland.c | 170 const uint16x8_t va3x0 = vbicq_u16(va3, vm0); in xnn_bf16_gemm_minmax_ukernel_4x4c8__neonfma_shland() local
|
| D | 4x4c8-minmax-neonfma-zip.c | 170 const uint16x8_t va3x0 = vbicq_u16(va3, vm0); in xnn_bf16_gemm_minmax_ukernel_4x4c8__neonfma_zip() local
|
| D | 5x4c8-minmax-neonfma-zip.c | 192 const uint16x8_t va3x0 = vbicq_u16(va3, vm0); in xnn_bf16_gemm_minmax_ukernel_5x4c8__neonfma_zip() local
|
| D | 5x4c8-minmax-neonfma-shland.c | 192 const uint16x8_t va3x0 = vbicq_u16(va3, vm0); in xnn_bf16_gemm_minmax_ukernel_5x4c8__neonfma_shland() local
|