| /external/XNNPACK/src/f32-dwconv2d-chw/gen/ |
| D | 3x3p1-minmax-neonfma-1x4.c | 92 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vi1x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_1x4() local
140 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_1x4() local
|
| D | 3x3p1-minmax-wasmsimd-arm-splat-1x4.c | 95 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_1x4() local
141 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_1x4() local
|
| D | 3x3p1-minmax-wasmsimd-x86-splat-1x4.c | 95 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_1x4() local
141 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_1x4() local
|
| D | 3x3p1-minmax-neon-1x4.c | 92 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vi1x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_1x4() local
140 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_1x4() local
|
| D | 3x3p1-minmax-wasmsimd-x86-splat-1x4-acc2.c | 95 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_1x4_acc2() local
142 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_1x4_acc2() local
|
| D | 3x3p1-minmax-neonfma-1x4-acc2.c | 92 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vi1x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_1x4_acc2() local
141 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_1x4_acc2() local
|
| D | 3x3p1-minmax-wasmsimd-arm-splat-1x4-acc2.c | 95 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_1x4_acc2() local
142 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_1x4_acc2() local
|
| D | 3x3p1-minmax-neon-1x4-acc2.c | 92 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vi1x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_1x4_acc2() local
141 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_1x4_acc2() local
|
| D | 3x3p1-minmax-neon-1x4-acc3.c | 92 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vi1x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_1x4_acc3() local
142 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_1x4_acc3() local
|
| D | 3x3p1-minmax-wasmsimd-x86-splat-1x4-acc3.c | 95 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_1x4_acc3() local
143 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_1x4_acc3() local
|
| D | 3x3p1-minmax-neonfma-1x4-acc3.c | 92 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vi1x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_1x4_acc3() local
142 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_1x4_acc3() local
|
| D | 3x3p1-minmax-wasmsimd-arm-splat-1x4-acc3.c | 95 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_1x4_acc3() local
143 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_1x4_acc3() local
|
| D | 3x3p1-minmax-wasmsimd-x86-splat-1x4-acc4.c | 95 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_1x4_acc4() local
144 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_1x4_acc4() local
|
| D | 3x3p1-minmax-neon-1x4-acc4.c | 92 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vi1x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_1x4_acc4() local
143 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_1x4_acc4() local
|
| D | 3x3p1-minmax-ssse3-1x4.c | 99 …const __m128 vi1x5678 = _mm_castsi128_ps(_mm_alignr_epi8(_mm_castps_si128(vi1x89AB), _mm_castps_si… in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_1x4() local
140 … const __m128 vi1x5678 = _mm_castsi128_ps(_mm_alignr_epi8(vzero, _mm_castps_si128(vi1x4567), 4)); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_1x4() local
|
| D | 3x3p1-minmax-neonfma-1x4-acc4.c | 92 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vi1x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_1x4_acc4() local
143 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_1x4_acc4() local
|
| D | 3x3p1-minmax-wasmsimd-arm-splat-1x4-acc4.c | 95 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_1x4_acc4() local
144 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_1x4_acc4() local
|
| D | 3x3p1-minmax-ssse3-1x4-acc3.c | 99 …const __m128 vi1x5678 = _mm_castsi128_ps(_mm_alignr_epi8(_mm_castps_si128(vi1x89AB), _mm_castps_si… in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_1x4_acc3() local
142 … const __m128 vi1x5678 = _mm_castsi128_ps(_mm_alignr_epi8(vzero, _mm_castps_si128(vi1x4567), 4)); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_1x4_acc3() local
|
| D | 3x3p1-minmax-wasmsimd-arm-loadsplat-1x4.c | 106 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_loadsplat_1x4() local
145 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_loadsplat_1x4() local
|
| D | 3x3p1-minmax-wasmsimd-x86-loadsplat-1x4.c | 106 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_loadsplat_1x4() local
145 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_loadsplat_1x4() local
|
| D | 3x3p1-minmax-ssse3-1x4-acc2.c | 99 …const __m128 vi1x5678 = _mm_castsi128_ps(_mm_alignr_epi8(_mm_castps_si128(vi1x89AB), _mm_castps_si… in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_1x4_acc2() local
141 … const __m128 vi1x5678 = _mm_castsi128_ps(_mm_alignr_epi8(vzero, _mm_castps_si128(vi1x4567), 4)); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_1x4_acc2() local
|
| D | 3x3p1-minmax-wasmsimd-arm-loadsplat-1x4-acc2.c | 106 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_loadsplat_1x4_acc2() local
146 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_loadsplat_1x4_acc2() local
|
| D | 3x3p1-minmax-ssse3-1x4-acc4.c | 99 …const __m128 vi1x5678 = _mm_castsi128_ps(_mm_alignr_epi8(_mm_castps_si128(vi1x89AB), _mm_castps_si… in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_1x4_acc4() local
143 … const __m128 vi1x5678 = _mm_castsi128_ps(_mm_alignr_epi8(vzero, _mm_castps_si128(vi1x4567), 4)); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_1x4_acc4() local
|
| D | 3x3p1-minmax-wasmsimd-x86-loadsplat-1x4-acc3.c | 106 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vi1x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_loadsplat_1x4_acc3() local
147 const v128_t vi1x5678 = wasm_v32x4_shuffle(vi1x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_loadsplat_1x4_acc3() local
|
| D | 3x3p1-minmax-neon-2x4.c | 110 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vi1x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_2x4() local
175 const float32x4_t vi1x5678 = vextq_f32(vi1x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_2x4() local
|