We might want that specifically for gdx as generating sboms specifically on aarch64 is really slow for GDX.