在一个集成电路中,多个芯片、共享内存以及互连形成了一个紧密集成的多处理核心(参见图4)。
On a single integrated circuit, multiple chips, Shared memory, and an interconnect form a tightly integrated core for multiprocessing (see Figure 4).
每个pod具有自己的处理器和内存,并通过一条高速缓存一致性互连总线(cache - coherent interconnect bus)连接到较大的系统。
Each pod has its own processors and memory, and is connected to the larger system through a cache-coherent interconnect bus.
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。
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.
应用推荐