英伟达Vera Rubin NVL72系统在MLPerf推理v6.1基准测试中实现了领先性能，展示了AI推理效率的进步。

System performance, efficient infrastructure scaling and continuous software optimization are key levers that determine AI inference economics. Higher system performance means more tokens generated, resulting in higher revenue. Efficient scaling means throughput grows proportionally as hardware gets added, requiring fewer resources to serve users at scale. Continuous optimization means generating more value from infr...