Koji Kai

Optimizing the DRAM refresh count for merged DRAM/logic LSIs

In merged DRAM/logic LSIs, the DRAM portion could suffer from shorter data retention time because of heat and noise caused by the logic portion. Frequent refreshes increase power consumption. Also, they disturb normal DRAM accesses leading to performance degradation. In order to overcome this problem, we propose several DRAM refresh architectures. The basic idea is to eliminate unnecessary DRAM refreshes. We have estimated the DRAM refresh count in executing benchmark programs under several architecture models. As a result, in the most effective combination of the architectures, we have obtained more than 80% reduction against a conventional DRAM refresh architecture for most benchmark programs. In addition to it, even when we have taken normal DRAM access into account, we have obtained more than 50% reduction for several benchmarks.

Performance/Energy Efficiency of Variable Line-Size Caches for Intelligent Memory Systems

Integrating main memory (DRAM) and processors into a single chip, or merged DRAM/logic LSI, makes it possible to exploit high on-chip memory bandwidth by widening on-chip bus and on-chip DRAM array. In addition, from energy point of view, the integration brings a significant improvement by decreasing the number of off-chip accesses.