/external/XNNPACK/src/f32-dwconv2d-chw/gen/ |
D | 3x3s2p1-minmax-wasmsimd-x86-loadsplat-2x4.c | 109 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4() local 120 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4() 121 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4() 185 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4() local 195 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4() 196 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4()
|
D | 3x3s2p1-minmax-wasmsimd-x86-loadsplat-2x4-acc2.c | 109 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4_acc2() local 120 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4_acc2() 121 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4_acc2() 187 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4_acc2() local 197 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4_acc2() 198 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_2x4_acc2()
|
D | 3x3s2p1-minmax-wasmsimd-arm-loadsplat-2x4-acc2.c | 109 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4_acc2() local 120 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4_acc2() 121 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4_acc2() 187 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4_acc2() local 197 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4_acc2() 198 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4_acc2()
|
D | 3x3s2p1-minmax-wasmsimd-arm-splat-2x4-acc2.c | 99 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4_acc2() local 110 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4_acc2() 111 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4_acc2() 177 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4_acc2() local 187 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4_acc2() 188 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4_acc2()
|
D | 3x3s2p1-minmax-wasmsimd-arm-splat-2x4.c | 99 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4() local 110 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4() 111 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4() 175 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4() local 185 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4() 186 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_2x4()
|
D | 3x3s2p1-minmax-wasmsimd-arm-loadsplat-2x4.c | 109 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4() local 120 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4() 121 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4() 185 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4() local 195 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4() 196 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_2x4()
|
D | 3x3s2p1-minmax-sse-2x4-acc2.c | 100 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4_acc2() local 111 const __m128 vi4x8ACE = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0, 2, 0)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4_acc2() 112 const __m128 vi4x9BDF = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1, 3, 1)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4_acc2() 179 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4_acc2() local 189 …const __m128 vi4x8ACE = _mm_and_ps(vmask_even, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4_acc2() 190 …const __m128 vi4x9BDF = _mm_and_ps(vmask_odd, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4_acc2()
|
D | 3x3s2p1-minmax-sse-2x4.c | 100 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4() local 111 const __m128 vi4x8ACE = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0, 2, 0)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4() 112 const __m128 vi4x9BDF = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1, 3, 1)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4() 177 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4() local 187 …const __m128 vi4x8ACE = _mm_and_ps(vmask_even, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4() 188 …const __m128 vi4x9BDF = _mm_and_ps(vmask_odd, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_2x4()
|
D | 3x3s2p1-minmax-wasmsimd-x86-splat-2x4.c | 99 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4() local 110 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4() 111 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4() 175 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4() local 185 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4() 186 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4()
|
D | 3x3s2p1-minmax-wasmsimd-x86-splat-2x4-acc2.c | 99 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4_acc2() local 110 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4_acc2() 111 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4_acc2() 177 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4_acc2() local 187 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4_acc2() 188 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_2x4_acc2()
|
D | 3x3s2p1-minmax-wasmsimd-x86-loadsplat-3x4.c | 122 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_3x4() local 139 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_3x4() 140 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_3x4() 225 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_3x4() local 239 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_3x4() 240 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_3x4()
|
D | 3x3s2p1-minmax-wasmsimd-arm-loadsplat-3x4.c | 122 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_3x4() local 139 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_3x4() 140 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_3x4() 225 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_3x4() local 239 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_3x4() 240 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_3x4()
|
D | 3x3s2p1-minmax-wasmsimd-arm-splat-3x4.c | 112 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_3x4() local 129 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_3x4() 130 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_3x4() 215 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_3x4() local 229 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_3x4() 230 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_3x4()
|
D | 3x3s2p1-minmax-wasmsimd-x86-splat-3x4.c | 112 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_3x4() local 129 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_3x4() 130 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_3x4() 215 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_3x4() local 229 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_3x4() 230 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_3x4()
|
D | 3x3s2p1-minmax-sse-3x4.c | 112 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_3x4() local 129 const __m128 vi4x8ACE = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0, 2, 0)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_3x4() 130 const __m128 vi4x9BDF = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1, 3, 1)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_3x4() 218 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_3x4() local 232 …const __m128 vi4x8ACE = _mm_and_ps(vmask_even, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_3x4() 233 …const __m128 vi4x9BDF = _mm_and_ps(vmask_odd, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_3x4()
|
D | 3x3s2p1-minmax-wasmsimd-arm-loadsplat-4x4.c | 135 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_4x4() local 158 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_4x4() 159 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_4x4() 265 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_4x4() local 283 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_4x4() 284 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_loadsplat_4x4()
|
D | 3x3s2p1-minmax-wasmsimd-x86-loadsplat-4x4.c | 135 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_4x4() local 158 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_4x4() 159 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_4x4() 265 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_4x4() local 283 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_4x4() 284 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_loadsplat_4x4()
|
D | 3x3s2p1-minmax-wasmsimd-arm-splat-4x4.c | 125 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_4x4() local 148 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_4x4() 149 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_4x4() 255 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_4x4() local 273 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_4x4() 274 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_arm_splat_4x4()
|
D | 3x3s2p1-minmax-wasmsimd-x86-splat-4x4.c | 125 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_4x4() local 148 const v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_4x4() 149 const v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_4x4() 255 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_4x4() local 273 …const v128_t vi4x8ACE = wasm_v128_and(vmask_even, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_4x4() 274 …const v128_t vi4x9BDF = wasm_v128_and(vmask_odd, wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, … in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__wasmsimd_x86_splat_4x4()
|
D | 3x3s2p1-minmax-sse-4x4.c | 124 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_4x4() local 147 const __m128 vi4x8ACE = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0, 2, 0)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_4x4() 148 const __m128 vi4x9BDF = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1, 3, 1)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_4x4() 259 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_4x4() local 277 …const __m128 vi4x8ACE = _mm_and_ps(vmask_even, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_4x4() 278 …const __m128 vi4x9BDF = _mm_and_ps(vmask_odd, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_4x4()
|
D | 3x3s2p1-minmax-sse-5x4.c | 136 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_5x4() local 165 const __m128 vi4x8ACE = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0, 2, 0)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_5x4() 166 const __m128 vi4x9BDF = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1, 3, 1)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_5x4() 300 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_5x4() local 322 …const __m128 vi4x8ACE = _mm_and_ps(vmask_even, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_5x4() 323 …const __m128 vi4x9BDF = _mm_and_ps(vmask_odd, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_5x4()
|
D | 3x3s2p1-minmax-sse-6x4.c | 148 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_6x4() local 183 const __m128 vi4x8ACE = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0, 2, 0)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_6x4() 184 const __m128 vi4x9BDF = _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1, 3, 1)); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_6x4() 341 const __m128 vi4xCDEF = _mm_loadu_ps(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_6x4() local 367 …const __m128 vi4x8ACE = _mm_and_ps(vmask_even, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(2, 0… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_6x4() 368 …const __m128 vi4x9BDF = _mm_and_ps(vmask_odd, _mm_shuffle_ps(vi4x89AB, vi4xCDEF, _MM_SHUFFLE(3, 1… in xnn_f32_dwconv2d_chw_ukernel_3x3s2p1__sse_6x4()
|
D | 5x5s2p2-minmax-wasmsimd-arm-loadsplat-1x4-acc3.c | 120 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_5x5s2p2__wasmsimd_arm_loadsplat_1x4_acc3() local 131 v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_5x5s2p2__wasmsimd_arm_loadsplat_1x4_acc3() 132 v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_5x5s2p2__wasmsimd_arm_loadsplat_1x4_acc3()
|
D | 5x5s2p2-minmax-wasmsimd-arm-loadsplat-1x4-acc2.c | 120 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_5x5s2p2__wasmsimd_arm_loadsplat_1x4_acc2() local 131 v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_5x5s2p2__wasmsimd_arm_loadsplat_1x4_acc2() 132 v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_5x5s2p2__wasmsimd_arm_loadsplat_1x4_acc2()
|
D | 5x5s2p2-minmax-wasmsimd-x86-loadsplat-1x4.c | 120 const v128_t vi4xCDEF = wasm_v128_load(i4 + 4); in xnn_f32_dwconv2d_chw_ukernel_5x5s2p2__wasmsimd_x86_loadsplat_1x4() local 131 v128_t vi4x8ACE = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 0, 2, 4, 6); in xnn_f32_dwconv2d_chw_ukernel_5x5s2p2__wasmsimd_x86_loadsplat_1x4() 132 v128_t vi4x9BDF = wasm_v32x4_shuffle(vi4x89AB, vi4xCDEF, 1, 3, 5, 7); in xnn_f32_dwconv2d_chw_ukernel_5x5s2p2__wasmsimd_x86_loadsplat_1x4()
|