40   const v128_t vw89AB = wasm_v128_load(weights + 8);  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()  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_x86_splat_3x4_acc2()
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_x86_splat_3x4_acc2()
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_x86_splat_3x4_acc2()
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_x86_splat_3x4_acc2()
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_x86_splat_3x4_acc2()
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_x86_splat_3x4_acc2()
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_x86_splat_3x4_acc2()
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_x86_splat_3x4_acc2()
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_x86_splat_3x4_acc2()
239 …vo0p1 = wasm_f32x4_add(vo0p1, wasm_f32x4_mul(vi1x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()
240 …vo1p1 = wasm_f32x4_add(vo1p1, wasm_f32x4_mul(vi2x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()
241 …vo2p1 = wasm_f32x4_add(vo2p1, wasm_f32x4_mul(vi3x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()
296 …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_x86_splat_3x4_acc2()
297 …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_x86_splat_3x4_acc2()
298 …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_x86_splat_3x4_acc2()
363 …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_x86_splat_3x4_acc2()
364 …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_x86_splat_3x4_acc2()
365 …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_x86_splat_3x4_acc2()
387 …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_x86_splat_3x4_acc2()
388 …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_x86_splat_3x4_acc2()
389 …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_x86_splat_3x4_acc2()
422 …vo0p1 = wasm_f32x4_add(vo0p1, wasm_f32x4_mul(vi1x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()
423 …vo1p1 = wasm_f32x4_add(vo1p1, wasm_f32x4_mul(vi2x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()
424 …vo2p1 = wasm_f32x4_add(vo2p1, wasm_f32x4_mul(vi3x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()
474 …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_x86_splat_3x4_acc2()
475 …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_x86_splat_3x4_acc2()
476 …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_x86_splat_3x4_acc2()
534 …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_x86_splat_3x4_acc2()
535 …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_x86_splat_3x4_acc2()
536 …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_x86_splat_3x4_acc2()
558 …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_x86_splat_3x4_acc2()
559 …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_x86_splat_3x4_acc2()
560 …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_x86_splat_3x4_acc2()
586 …vo0p1 = wasm_f32x4_add(vo0p1, wasm_f32x4_mul(vi1x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()
587 …vo1p1 = wasm_f32x4_add(vo1p1, wasm_f32x4_mul(vi2x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()
588 …vo2p1 = wasm_f32x4_add(vo2p1, wasm_f32x4_mul(vi3x6789, wasm_v32x4_shuffle(vw89AB, vw89AB, 2, 2, 2,…  in xnn_f32_dwconv2d_chw_ukernel_5x5p2__wasmsimd_x86_splat_3x4_acc2()