/external/XNNPACK/src/f32-dwconv2d-chw/gen/ |
D | 5x5p2-minmax-neon-4x4.c | 265 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_4x4() local 483 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_4x4() local 681 const float32x4_t vi7x6789 = vextq_f32(vi7x5678, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_4x4() local
|
D | 5x5p2-minmax-wasmsimd-x86-splat-4x4.c | 266 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_4x4() local 483 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_4x4() local 680 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_4x4() local
|
D | 5x5p2-minmax-neonfma-4x4.c | 265 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_4x4() local 483 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_4x4() local 681 const float32x4_t vi7x6789 = vextq_f32(vi7x5678, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_4x4() local
|
D | 5x5p2-minmax-wasmsimd-arm-splat-4x4.c | 266 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_4x4() local 483 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_4x4() local 680 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_4x4() local
|
D | 5x5p2-minmax-wasmsimd-arm-splat-4x4-acc2.c | 266 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_4x4_acc2() local 487 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_4x4_acc2() local 688 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_4x4_acc2() local
|
D | 5x5p2-minmax-wasmsimd-x86-splat-4x4-acc2.c | 266 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_4x4_acc2() local 487 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_4x4_acc2() local 688 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_4x4_acc2() local
|
D | 5x5p2-minmax-wasmsimd-x86-splat-5x4.c | 300 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_5x4() local 554 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_5x4() local 786 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_5x4() local
|
D | 5x5p2-minmax-wasmsimd-x86-loadsplat-4x4.c | 292 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_loadsplat_4x4() local 509 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_loadsplat_4x4() local 706 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_loadsplat_4x4() local
|
D | 5x5p2-minmax-neon-4x4-acc2.c | 265 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_4x4_acc2() local 487 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_4x4_acc2() local 689 const float32x4_t vi7x6789 = vextq_f32(vi7x5678, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_4x4_acc2() local
|
D | 5x5p2-minmax-wasmsimd-arm-loadsplat-4x4.c | 292 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_loadsplat_4x4() local 509 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_loadsplat_4x4() local 706 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_loadsplat_4x4() local
|
D | 5x5p2-minmax-wasmsimd-arm-splat-5x4.c | 300 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_5x4() local 554 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_5x4() local 786 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_5x4() local
|
D | 5x5p2-minmax-neonfma-4x4-acc2.c | 265 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_4x4_acc2() local 487 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_4x4_acc2() local 689 const float32x4_t vi7x6789 = vextq_f32(vi7x5678, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_4x4_acc2() local
|
D | 5x5p2-minmax-wasmsimd-arm-loadsplat-4x4-acc2.c | 292 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_loadsplat_4x4_acc2() local 513 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_loadsplat_4x4_acc2() local 714 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_loadsplat_4x4_acc2() local
|
D | 5x5p2-minmax-wasmsimd-x86-loadsplat-4x4-acc2.c | 292 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_loadsplat_4x4_acc2() local 513 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_loadsplat_4x4_acc2() local 714 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_loadsplat_4x4_acc2() local
|
D | 5x5p2-minmax-neon-5x4.c | 299 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_5x4() local 554 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_5x4() local 787 const float32x4_t vi7x6789 = vextq_f32(vi7x5678, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neon_5x4() local
|
D | 5x5p2-minmax-neonfma-5x4.c | 299 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_5x4() local 554 const float32x4_t vi7x6789 = vextq_f32(vi7x4567, vi7x89AB, 2); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_5x4() local 787 const float32x4_t vi7x6789 = vextq_f32(vi7x5678, vzero, 1); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__neonfma_5x4() local
|
D | 5x5p2-minmax-wasmsimd-x86-loadsplat-5x4.c | 326 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_loadsplat_5x4() local 580 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_loadsplat_5x4() local 812 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_loadsplat_5x4() local
|
D | 5x5p2-minmax-wasmsimd-arm-loadsplat-5x4.c | 326 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_loadsplat_5x4() local 580 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x4567, vi7x89AB, 2, 3, 4, 5); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_loadsplat_5x4() local 812 const v128_t vi7x6789 = wasm_v32x4_shuffle(vi7x5678, vzero, 1, 2, 3, 4); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_loadsplat_5x4() local
|
D | 5x5p2-minmax-sse-4x4.c | 298 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vi7x89AB, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_4x4() local 512 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vi7x89AB, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_4x4() local 706 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vzero, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_4x4() local
|
D | 5x5p2-minmax-sse-4x4-acc2.c | 298 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vi7x89AB, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_4x4_acc2() local 516 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vi7x89AB, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_4x4_acc2() local 714 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vzero, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_4x4_acc2() local
|
D | 5x5p2-minmax-sse-5x4.c | 336 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vi7x89AB, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_5x4() local 589 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vi7x89AB, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_5x4() local 819 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vzero, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_5x4() local
|
/external/XNNPACK/src/f16-dwconv2d-chw/gen/ |
D | 5x5p2-minmax-neonfp16arith-4x4.c | 263 const float16x4_t vi7x6789 = vext_f16(vi7x4567, vi7x89AB, 2); in xnn_f16_dwconv2d_chw_ukernel_5x5p2__neonfp16arith_4x4() local 481 const float16x4_t vi7x6789 = vext_f16(vi7x4567, vi7x89AB, 2); in xnn_f16_dwconv2d_chw_ukernel_5x5p2__neonfp16arith_4x4() local 679 const float16x4_t vi7x6789 = vext_f16(vi7x5678, vzero, 1); in xnn_f16_dwconv2d_chw_ukernel_5x5p2__neonfp16arith_4x4() local
|
D | 5x5p2-minmax-neonfp16arith-4x4-acc2.c | 263 const float16x4_t vi7x6789 = vext_f16(vi7x4567, vi7x89AB, 2); in xnn_f16_dwconv2d_chw_ukernel_5x5p2__neonfp16arith_4x4_acc2() local 485 const float16x4_t vi7x6789 = vext_f16(vi7x4567, vi7x89AB, 2); in xnn_f16_dwconv2d_chw_ukernel_5x5p2__neonfp16arith_4x4_acc2() local 687 const float16x4_t vi7x6789 = vext_f16(vi7x5678, vzero, 1); in xnn_f16_dwconv2d_chw_ukernel_5x5p2__neonfp16arith_4x4_acc2() local
|
D | 5x5p2-minmax-neonfp16arith-5x4.c | 297 const float16x4_t vi7x6789 = vext_f16(vi7x4567, vi7x89AB, 2); in xnn_f16_dwconv2d_chw_ukernel_5x5p2__neonfp16arith_5x4() local 552 const float16x4_t vi7x6789 = vext_f16(vi7x4567, vi7x89AB, 2); in xnn_f16_dwconv2d_chw_ukernel_5x5p2__neonfp16arith_5x4() local 785 const float16x4_t vi7x6789 = vext_f16(vi7x5678, vzero, 1); in xnn_f16_dwconv2d_chw_ukernel_5x5p2__neonfp16arith_5x4() local
|
/external/XNNPACK/src/amalgam/ |
D | sse.c | 3179 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vi7x89AB, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_4x4() local 3393 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vi7x89AB, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_4x4() local 3587 const __m128 vi7x6789 = _mm_shuffle_ps(vi7x5678, vzero, _MM_SHUFFLE(1, 0, 2, 1)); in xnn_f32_dwconv2d_chw_ukernel_5x5p2__sse_4x4() local
|