NoC-based Chip-Multiprocessors(CMPs) are promising mechanisms which are essential to satisfy a growing need for easily scalable and high-performance in recent decades. However, in the process of running the programs, there are some data that misses from LI cache to L2 cache. These cache misses have some effects on network congestion and access time. In this paper, we use the average number of hops to describe the fairness of the thread resources and explain each thread's access to each bank is fair. After then, we described an address remapping algorithm for cache miss in NoC-based CMPs. We evaluate our mechanism using Simplescalar and Cacti power simulation tool. Experimental results show that address remapping achieves significant improvement in the system performance and power consumption.
[1]
Leszek Koszalka,et al.
Evaluation Scheme for NoC based CMP with Integrated Processor Management System
,
2010
.
[2]
Huawei Li,et al.
Address Remapping for Static NUCA in NoC-Based Degradable Chip-Multiprocessors
,
2010,
2010 IEEE 16th Pacific Rim International Symposium on Dependable Computing.
[3]
Norman P. Jouppi,et al.
Improving direct-mapped cache performance by the addition of a small fully-associative cache and pre
,
1990,
ISCA 1990.
[4]
Jaehyuk Huh,et al.
A NUCA Substrate for Flexible CMP Cache Sharing
,
2007,
IEEE Transactions on Parallel and Distributed Systems.