Title :
Power-Efficient DRAM Speculation
Author :
Aggarwal, Nidhi ; Cantin, Jason F. ; Lipasti, Mikko H. ; Smith, James E.
Author_Institution :
Electr. & Comput. Eng. Dept., Univ. of Wisconsin, Madison, WI
Abstract :
Power-efficient DRAM Speculation (PEDS) is a power optimization targeted at broadcast-based shared-memory multiprocessor systems that speculatively access DRAM in parallel with the broadcast snoop. Although speculatively accessing DRAM has the potential performance advantage of overlapping DRAM latency with the snoop, it wastes power for memory requests that obtain data from other processorspsila caches. PEDS takes advantage of information provided by a Region Coherence Array to identify requests that have a high likelihood of obtaining data from another processorpsilas cache, and does not access DRAM speculatively for those requests. By doing so, PEDS eliminates DRAM reads, reduces DRAM power consumption, reduces contention for DRAM resources, and increases the opportunity for DRAM power management. PEDS requires almost no additional hardware in systems that incorporate Region Coherence Arrays. Detailed simulation results show PEDS reduces average DRAM read traffic 28-32%, reduces average DRAM power dissipation 17-22%, and reduces average DRAM energy consumption 16-21%.
Keywords :
DRAM chips; low-power electronics; power aware computing; shared memory systems; DRAM latency; DRAM power consumption; DRAM power management; broadcast snoop; broadcast-based shared-memory multiprocessor systems; power optimization; power-efficient DRAM speculation; region coherence arrays; Broadcasting; Content management; Delay; Energy consumption; Energy management; Hardware; Multiprocessing systems; Power system management; Random access memory; Resource management;
Conference_Titel :
High Performance Computer Architecture, 2008. HPCA 2008. IEEE 14th International Symposium on
Conference_Location :
Salt Lake City, UT
Print_ISBN :
978-1-4244-2070-4
DOI :
10.1109/HPCA.2008.4658649