/external/XNNPACK/src/f32-dwconv2d-chw/gen/ |
D | 5x5p2-minmax-neonfma-1x4-acc2.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc2() local 245 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc2() local 346 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc2() local
|
D | 5x5p2-minmax-neon-1x4.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4() local 244 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4() local 344 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4() local
|
D | 5x5p2-minmax-wasmsimd-x86-splat-1x4.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4() local 246 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4() local 344 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4() local
|
D | 5x5p2-minmax-wasmsimd-arm-splat-1x4-acc2.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc2() local 247 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc2() local 346 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc2() local
|
D | 5x5p2-minmax-neonfma-1x4.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4() local 244 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4() local 344 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4() local
|
D | 5x5p2-minmax-neon-1x4-acc2.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc2() local 245 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc2() local 346 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc2() local
|
D | 5x5p2-minmax-wasmsimd-arm-splat-1x4.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4() local 246 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4() local 344 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4() local
|
D | 5x5p2-minmax-wasmsimd-x86-splat-1x4-acc2.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc2() local 247 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc2() local 346 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc2() local
|
D | 5x5p2-minmax-neonfma-1x4-acc3.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc3() local 246 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc3() local 348 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc3() local
|
D | 5x5p2-minmax-wasmsimd-x86-splat-1x4-acc4.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc4() local 249 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc4() local 350 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc4() local
|
D | 5x5p2-minmax-neonfma-1x4-acc4.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc4() local 247 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc4() local 350 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc4() local
|
D | 5x5p2-minmax-wasmsimd-arm-splat-1x4-acc4.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc4() local 249 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc4() local 350 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc4() local
|
D | 5x5p2-minmax-wasmsimd-arm-splat-1x4-acc5.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc5() local 250 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc5() local 352 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc5() local
|
D | 5x5p2-minmax-neon-1x4-acc3.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc3() local 246 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc3() local 348 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc3() local
|
D | 5x5p2-minmax-wasmsimd-x86-splat-1x4-acc3.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc3() local 248 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc3() local 348 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc3() local
|
D | 5x5p2-minmax-neon-1x4-acc4.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc4() local 247 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc4() local 350 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc4() local
|
D | 5x5p2-minmax-wasmsimd-arm-splat-1x4-acc3.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc3() local 248 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc3() local 348 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_1x4_acc3() local
|
D | 5x5p2-minmax-wasmsimd-x86-splat-1x4-acc5.c | 140 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc5() local 250 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc5() local 352 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_1x4_acc5() local
|
D | 3x3p1-minmax-wasmsimd-x86-splat-3x4.c | 134 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_3x4() local 214 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_x86_splat_3x4() local
|
D | 3x3p1-minmax-wasmsimd-arm-splat-3x4.c | 134 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vi4x89AB, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_3x4() local 214 const v128_t vi4x5678 = wasm_v32x4_shuffle(vi4x4567, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__wasmsimd_arm_splat_3x4() local
|
D | 3x3p1-minmax-neonfma-3x4.c | 131 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_3x4() local 213 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neonfma_3x4() local
|
D | 3x3p1-minmax-neon-3x4.c | 131 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_3x4() local 213 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__neon_3x4() local
|
D | 5x5p2-minmax-neonfma-1x4-acc5.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc5() local 248 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc5() local 352 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_1x4_acc5() local
|
D | 5x5p2-minmax-neon-1x4-acc5.c | 137 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc5() local 248 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vi4x89AB, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc5() local 352 const float32x4_t vi4x5678 = vextq_f32(vi4x4567, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_1x4_acc5() local
|
D | 3x3p1-minmax-ssse3-3x4.c | 140 …const __m128 vi4x5678 = _mm_castsi128_ps(_mm_alignr_epi8(_mm_castps_si128(vi4x89AB), _mm_castps_si… in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_3x4() local 215 … const __m128 vi4x5678 = _mm_castsi128_ps(_mm_alignr_epi8(vzero, _mm_castps_si128(vi4x4567), 4)); in xnn_f32_dwconv2d_chw_ukernel_3x3p1__ssse3_3x4() local
|