⚡ 本页包含 AI 生成的分析内容,仅供参考
本文介绍了一款用于45nm 8核企业级Xeon处理器的片上环形互连架构,实现了1.2 TB/s的高带宽和低延迟共享L3缓存访问,同时兼顾设计模块化、低功耗和低面积开销。
Xeon microprocessors targeting high performance, low-power products [1, 2]. Several key design requirements for this product include high bandwidth, low-latency shared L3 cache access, design modularity to support efficient release of multiple products, and low power and silicon area overhead. A ring interconnect is particularly well suited to achieve all of these design requirements [3]. The implementation details for accomplishing the design target will be described in this paper. Figure 9.4.1 is a floorplan view of the chip design showing the relative placement of the 10 ring stops (Cbox, Sbox), the co-located cores and cache (LLC), and the ring interconnect. The ring interconnect consists of two sets of physical channels that communicate in opposite directions, clock-wise (CW) and counterclock-wise (CCW). Each ring stop can independently handle ring traffic fr
Cheolmin Park, Roy Badeau, Larry Biro, Jonathan Chang,
Tejpal Singh, Jim Vash, Bo Wang, Tom Wang Intel, Hudson, MA Nehalem-EX® is the most recent product in the family of the X86