Title :
pbCAM: Probabilistically-banked Content Addressable Memory
Author :
Soyata, Tolga ; Liobe, John
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Rochester, Rochester, NY, USA
Abstract :
Content Addressable Memories find wide use in network routers and certain image processing applications. However, their use is limited due to their high power demand resulting from their high activity factor. A banked CAM, on the other hand, partitions the entire CAM into smaller banks to cut down on excessive searches, while it may reduce the effective CAM size when the data entries are unevenly distributed. A new banked CAM design is introduced in this paper which achieves significant energy and power savings through the use of Bloom Filters by effectively decoupling data elements from their bank index. Simulation results show energy savings of nearly an order of magnitude.
Keywords :
content-addressable storage; data structures; image processing; network routing; Bloom filters; activity factor; bank index; banked CAM; decoupling data elements; image processing; network routers; pbCAM; power demand; probabilistically-banked content addressable memory; Calculators; Computer aided manufacturing; Energy consumption; Equations; Mathematical model; Power demand; Transistors;
Conference_Titel :
SOC Conference (SOCC), 2012 IEEE International
Conference_Location :
Niagara Falls, NY
Print_ISBN :
978-1-4673-1294-3
DOI :
10.1109/SOCC.2012.6398372