/external/XNNPACK/src/f32-dwconv2d-chw/gen/ |
D | 3x3s2p1-minmax-neon-3x4.c | 126 const float32x4_t vi5x7BDF = vextq_f32(vi5x1357, vi5x8ACE9BDF.val[1], 3); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__neon_3x4() local 215 const float32x4_t vi5x7BDF = vextq_f32(vi5x1357, vi5x9BDF, 3); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__neon_3x4() local
|
D | 3x3s2p1-minmax-neonfma-3x4.c | 126 const float32x4_t vi5x7BDF = vextq_f32(vi5x1357, vi5x8ACE9BDF.val[1], 3); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__neonfma_3x4() local 215 const float32x4_t vi5x7BDF = vextq_f32(vi5x1357, vi5x9BDF, 3); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__neonfma_3x4() local
|
D | 3x3s2p1-minmax-neon-4x4.c | 144 const float32x4_t vi5x7BDF = vextq_f32(vi5x1357, vi5x8ACE9BDF.val[1], 3); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__neon_4x4() local 256 const float32x4_t vi5x7BDF = vextq_f32(vi5x1357, vi5x9BDF, 3); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__neon_4x4() local
|
D | 3x3s2p1-minmax-neonfma-4x4.c | 144 const float32x4_t vi5x7BDF = vextq_f32(vi5x1357, vi5x8ACE9BDF.val[1], 3); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__neonfma_4x4() local 256 const float32x4_t vi5x7BDF = vextq_f32(vi5x1357, vi5x9BDF, 3); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__neonfma_4x4() local
|
D | 3x3s2p1-minmax-wasmsimd-x86-splat-3x4.c | 156 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_3x4() local 251 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_3x4() local
|
D | 3x3s2p1-minmax-wasmsimd-arm-splat-3x4.c | 156 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_3x4() local 251 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_3x4() local
|
D | 3x3s2p1-minmax-wasmsimd-arm-loadsplat-3x4.c | 166 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_3x4() local 261 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_3x4() local
|
D | 3x3s2p1-minmax-wasmsimd-x86-loadsplat-3x4.c | 166 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_3x4() local 261 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_3x4() local
|
D | 3x3s2p1-minmax-sse-3x4.c | 169 const __m128 vi5x7BDF = _mm_move_ss(vi5xF9BD, vi5x7531); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_3x4() local 272 const __m128 vi5x7BDF = _mm_move_ss(vi5xF9BD, vi5x7531); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_3x4() local
|
D | 3x3s2p1-minmax-wasmsimd-x86-splat-4x4.c | 182 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_4x4() local 302 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_4x4() local
|
D | 3x3s2p1-minmax-wasmsimd-arm-splat-4x4.c | 182 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_4x4() local 302 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_4x4() local
|
D | 3x3s2p1-minmax-wasmsimd-arm-loadsplat-4x4.c | 192 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_4x4() local 312 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_4x4() local
|
D | 3x3s2p1-minmax-wasmsimd-x86-loadsplat-4x4.c | 192 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_4x4() local 312 const v128_t vi5x7BDF = wasm_v32x4_shuffle(vi5x1357, vi5x9BDF, 3, 4, 5, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_4x4() local
|
D | 3x3s2p1-minmax-sse-4x4.c | 199 const __m128 vi5x7BDF = _mm_move_ss(vi5xF9BD, vi5x7531); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_4x4() local 329 const __m128 vi5x7BDF = _mm_move_ss(vi5xF9BD, vi5x7531); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_4x4() local
|
D | 3x3s2p1-minmax-sse-5x4.c | 229 const __m128 vi5x7BDF = _mm_move_ss(vi5xF9BD, vi5x7531); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_5x4() local 386 const __m128 vi5x7BDF = _mm_move_ss(vi5xF9BD, vi5x7531); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_5x4() local
|
D | 3x3s2p1-minmax-sse-6x4.c | 259 const __m128 vi5x7BDF = _mm_move_ss(vi5xF9BD, vi5x7531); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_6x4() local 443 const __m128 vi5x7BDF = _mm_move_ss(vi5xF9BD, vi5x7531); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_6x4() local
|
/external/XNNPACK/src/f16-dwconv2d-chw/gen/ |
D | 3x3s2p1-minmax-neonfp16arith-3x4.c | 126 const float16x4_t vi5x7BDF = vext_f16(vi5x1357, vi5x8ACE9BDF.val[1], 3); in xnn_f16_dwconv2d_chw_ukernel_3x3s2p1__neonfp16arith_3x4() local 215 const float16x4_t vi5x7BDF = vext_f16(vi5x1357, vi5x9BDF, 3); in xnn_f16_dwconv2d_chw_ukernel_3x3s2p1__neonfp16arith_3x4() local
|
D | 3x3s2p1-minmax-neonfp16arith-4x4.c | 144 const float16x4_t vi5x7BDF = vext_f16(vi5x1357, vi5x8ACE9BDF.val[1], 3); in xnn_f16_dwconv2d_chw_ukernel_3x3s2p1__neonfp16arith_4x4() local 256 const float16x4_t vi5x7BDF = vext_f16(vi5x1357, vi5x9BDF, 3); in xnn_f16_dwconv2d_chw_ukernel_3x3s2p1__neonfp16arith_4x4() local
|