Data memory design considering effective bitwidth for low-energy embedded systems

Abstract

This paper presents a novel low-energy memory design technique, considering effective bitwidth of variables for applicationspecific systems, called VAbM technique. It targets the exploitation of both data locality and effective bitwidth of variables to reduce energy consumed by redundant bits. Under constraints of the number of memory banks, the VAbM technique use variable analysis results to perform allocating and assigning on-chip RAM into multiple memory banks, which have different size with different number of word lines and different number of bit lines tailored to each application requirements. Experimental results with several real embedded applications demonstrate significant energy reduction up to 64.8% over monolithic memory, and 18.4% over memory designed by banking technique.

Customized on-chip memories for embedded chip multiprocessors

“…DISCUSSION OF RELATED WORK Prior work on memory design for embedded systems focused on several aspects, including scratch-pad memory partitioning [1,16], cache partitioning [15], general memory design issues [3,13], and hierarchy design [4]. Meftali et al [9] considered the memory allocation problem, which they formulate based on an integer linear programming model.…”

Quality-driven design by bitwidth optimization for video applications

“…[8] discussed bitwidth analysis with application to silicon compilation and reduced silicon real estate by 15-86%, improved clock speed by 3-249%, and reduced power by 46-73%. [9] presented datapath optimization for power minimization and [10] presented low-energy memory design considering bitwidth. [6] discussed dynamically exploiting narrow width operands to improve processor power and performance.…”