Full threadesperent·What makes a 3d graphics rendering engine a good fit for llm inference?View on HN