← 返回 JSSC 论文列表JSSC 2013第8期Memory32nm
An Energy Efficient 32-nm 20-MB Shared On-Die L3 Cache for Intel® Xeon® Processor E5 Family Min Huang, Moty Mehalel, Ramesh Arvapalli, and
英特尔Xeon处理器E5系列中一款能效优化的32纳米20MB共享片上L3缓存设计
32nm工艺, 0.7V-1.0V电压范围, 1.2GHz-4.0GHz频率范围
L3缓存能效优化Xeon处理器高K金属栅工艺冗余设计
▸采用0.2119um高密度大阵列单元和0.2725um高性能小阵列单元
▸先进的节能方案和Vccmin设计技术提升能效
▸支持高密度模块化和能效优化的缓存拓扑结构
Abstract
An energy ef ficient on-die 20-way set associative L3 c a c h eo fs i z e2 0M Bf o rt h eI n t e l ®X e o n ®p r o c e s s o rE 5f a m i l yi s presented. It is manufactured in the Intel’s 32-nm second gener- ation of high-K dielectric metal gate process with 9-copper metal layers. The L3 cache design uses 0.2119 um cell for the high den- sity big array and 0.2725 um cell for the high performance smaller arrays. The power ef ficiency was achieved by employing advanced power saving schemes and effective Vccmin design techniques. The proposed L3 cache topology seamlessly supports a high density modular and energy efficient designs. The effective and rich redun- dancy design improves both yield and low voltage operations. The L3 cache achieves more than 20–40% energy ef ficiency when com- pared to previous generations and demonstrates wide operating ranges from 1.2 GHz at below 0.7 V to greater than 4.0 GHz at above 1.0 V.