NUMA alleviates these bottlenecks by limiting the number of CPUs on any one memory bus and connecting the various nodes by means of a high speed interconnection.

• NUMA通过限制任何一条内存总线上的CPU数量并依靠高速互连来连接各个节点,从而缓解了这些瓶颈状况。