Pseudo-Static Master-Slave Match-Line Scheme for Sustainable-Performance and Energy-Efficient Content Addressable Memory

Abstract

Content addressable memory (CAM) is used in applications requiring high-speed lookup operations. Search operations in CAM is prone to frequent switching of large capacitive match-line (ML) which causes huge energy dissipation. In this paper, we present an pseudo-static master-slave ML (PSMSML) architecture to increase the energy efficiency of MLs. In the proposed approach, the core cell employs a transmission-gate comparison and a single transistor evaluation while pseudo-static ML is partitioned into master and slave nodes to minimize charge variation. The proposed 64×32-bit macro achieves sustainable search-speed and dissipates only 0.96 fJ/bit/search under 1.1-V supply. Despite increasing energy-efficiency by 30%-44%, energy-delay-product also reduces around 9% and 29% in the proposed CAM when compared to a gated-power ML sensing and an existing high-performance MSML designs.