Leakage Energy Reduction in Cache Memory by Software Self-invalidation

Recently, energy dissipation by microprocessors is getting larger, which leads to a serious problem in terms of allowable temperature and performance improvement for future microprocessors. Cache memory is effective in bridging a growing speed gap between a processor and relatively slow external main memory, and has increased in its size. However, energy dissipation in the cache memory will approach or exceed 50% of the increasing total dissipation by processors. An important point to note is that, in the near future, static (leakage) energy will dominate the total energy consumption in deep sub-micron processes. In this paper, we propose cache memory architecture, especially for on-chip multiprocessors, that achieves efficient reduction of leakage energy in cache memories by exploiting gated-Vdd control and software self-invalidation. In the simulation, our technique reduced 46.5% of leakage energy at maximum, and 23.4% on average, in the execution of SPLASH-2 programs.

[1]  Bill Moyer,et al.  A low power unified cache architecture providing power and performance flexibility , 2000, ISLPED'00: Proceedings of the 2000 International Symposium on Low Power Electronics and Design (Cat. No.00TH8514).

[2]  Kaushik Roy,et al.  An integrated circuit/architecture approach to reducing leakage in deep-submicron high-performance I-caches , 2001, Proceedings HPCA Seventh International Symposium on High-Performance Computer Architecture.

[3]  Kevin Skadron,et al.  State-preserving vs. non-state-preserving leakage control in caches , 2004, Proceedings Design, Automation and Test in Europe Conference and Exhibition.

[4]  Anoop Gupta,et al.  The SPLASH-2 programs: characterization and methodological considerations , 1995, ISCA.

[5]  David Blaauw,et al.  Drowsy caches: simple techniques for reducing leakage power , 2002, ISCA.

[6]  David A. Wood,et al.  Dynamic self-invalidation: reducing coherence overhead in shared-memory multiprocessors , 1995, Proceedings 22nd Annual International Symposium on Computer Architecture.

[7]  Yan Meng,et al.  On the limits of leakage power reduction in caches , 2005, 11th International Symposium on High-Performance Computer Architecture.

[8]  Krste Asanovic,et al.  Dynamic fine-grain leakage reduction using leakage-biased bitlines , 2002, ISCA.

[9]  David L Weaver,et al.  The SPARC architecture manual : version 9 , 1994 .

[10]  Babak Falsafi,et al.  Selective, accurate, and timely self-invalidation using last-touch prediction , 2000, ISCA '00.

[11]  Koji Nii,et al.  A low power SRAM using auto-backgate-controlled MT-CMOS , 1998, Proceedings. 1998 International Symposium on Low Power Electronics and Design (IEEE Cat. No.98TH8379).

[12]  A. L. Rosenberg,et al.  Improving Replacement Decisions in Set-Associative Caches TITLE2: , 2001 .

[13]  Kaushik Roy,et al.  Gated-Vdd: a circuit technique to reduce leakage in deep-submicron cache memories , 2000, ISLPED '00.

[14]  Margaret Martonosi,et al.  Cache decay: exploiting generational behavior to reduce cache leakage power , 2001, ISCA 2001.