fix: use cute::crd2idx for SF remap — layout_sf() not directly callable

CuTe Layout objects with hierarchical shapes can't be called directly
with flat (m, k_sf). Use cute::crd2idx(make_coord(m, k_sf), layout_sf)
to convert logical coordinates to physical indices.

src/nvfp4_megamoe_kernel/cutlass_nvfp4_gemm/cutlass_nvfp4_gemm.cu

@@ -122,7 +122,7 @@ __global__ void remap_sf_to_cutlass_kernel(
     if (m >= MN || k_sf >= K_sf) return;
 
     // Compute the CUTLASS physical index from logical (m, k_sf) via the layout
-    int dst_idx = layout_sf(cute::make_coord(m, k_sf));
+    int dst_idx = cute::crd2idx(cute::make_coord(m, k_sf), layout_sf);
 
     // Read from source (row-major or col-major) and write to CUTLASS position
     dst[dst_idx] = col_major_src ? src[k_sf * MN + m] : src[m * K_sf + k_sf];