40   const v128_t vw89AB = wasm_v128_load(weights + 8);  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()  local
113 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi1x4567, wasm_v32x4_shuffle(vw89AB, vw89AB, 0, 0, 0,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
114 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi2x4567, wasm_v32x4_shuffle(vw89AB, vw89AB, 0, 0, 0,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
115 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi3x4567, wasm_v32x4_shuffle(vw89AB, vw89AB, 0, 0, 0,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
180 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi2x2345, wasm_v32x4_shuffle(vw89AB, vw89AB, 3, 3, 3,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
181 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi3x2345, wasm_v32x4_shuffle(vw89AB, vw89AB, 3, 3, 3,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
182 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi4x2345, wasm_v32x4_shuffle(vw89AB, vw89AB, 3, 3, 3,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
204 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi1x5678, wasm_v32x4_shuffle(vw89AB, vw89AB, 1, 1, 1,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
205 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi2x5678, wasm_v32x4_shuffle(vw89AB, vw89AB, 1, 1, 1,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
206 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi3x5678, wasm_v32x4_shuffle(vw89AB, vw89AB, 1, 1, 1,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
239 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi1x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
240 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi2x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
241 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi3x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
293 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi1x4567, wasm_v32x4_shuffle(vw89AB, vw89AB, 0, 0, 0,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
294 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi2x4567, wasm_v32x4_shuffle(vw89AB, vw89AB, 0, 0, 0,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
295 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi3x4567, wasm_v32x4_shuffle(vw89AB, vw89AB, 0, 0, 0,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
360 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi2x2345, wasm_v32x4_shuffle(vw89AB, vw89AB, 3, 3, 3,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
361 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi3x2345, wasm_v32x4_shuffle(vw89AB, vw89AB, 3, 3, 3,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
362 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi4x2345, wasm_v32x4_shuffle(vw89AB, vw89AB, 3, 3, 3,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
384 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi1x5678, wasm_v32x4_shuffle(vw89AB, vw89AB, 1, 1, 1,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
385 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi2x5678, wasm_v32x4_shuffle(vw89AB, vw89AB, 1, 1, 1,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
386 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi3x5678, wasm_v32x4_shuffle(vw89AB, vw89AB, 1, 1, 1,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
419 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi1x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
420 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi2x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
421 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi3x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
468 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi1x4567, wasm_v32x4_shuffle(vw89AB, vw89AB, 0, 0, 0,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
469 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi2x4567, wasm_v32x4_shuffle(vw89AB, vw89AB, 0, 0, 0,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
470 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi3x4567, wasm_v32x4_shuffle(vw89AB, vw89AB, 0, 0, 0,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
528 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi2x2345, wasm_v32x4_shuffle(vw89AB, vw89AB, 3, 3, 3,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
529 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi3x2345, wasm_v32x4_shuffle(vw89AB, vw89AB, 3, 3, 3,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
530 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi4x2345, wasm_v32x4_shuffle(vw89AB, vw89AB, 3, 3, 3,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
552 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi1x5678, wasm_v32x4_shuffle(vw89AB, vw89AB, 1, 1, 1,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
553 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi2x5678, wasm_v32x4_shuffle(vw89AB, vw89AB, 1, 1, 1,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
554 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi3x5678, wasm_v32x4_shuffle(vw89AB, vw89AB, 1, 1, 1,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
580 …vo0p0 = wasm_f32x4_add(vo0p0, wasm_f32x4_mul(vi1x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
581 …vo1p0 = wasm_f32x4_add(vo1p0, wasm_f32x4_mul(vi2x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()
582 …vo2p0 = wasm_f32x4_add(vo2p0, wasm_f32x4_mul(vi3x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_arm_splat_3x4()