Document Type

Technical Report

Publication Date

2003-10-03

Filename

wucse-2003-81.pdf

DOI:

10.7936/K7FT8JC4

Technical Report Number

WUCSE-2003-81

Abstract

We describe techniques for identifying and optimizing memory-accessing instruction sequences. We capture a sequence of such instructions, with the goal of sending the sequence as a single instruction from the CPU to a smart memory subsystem (IRAM or PIM). With a software/hardware codesign, the memory-accessing gestures can be rewritten as succinct superoperator instructions, and the gestures themselves could vary at runtime. As a result, the CPU executes fewer instructions and the CPU-memory bus is charged less often, resulting in lower power consumption. Reduction in power can be crucial for constrained, embedded systems. We discover gestures using a static and a dynamic approach, and we present data showing the presence of such gestures in real benchmarks (Java and C). We have shown the gesture-minimization problem to be NP-Complete, so we offer in this paper a heuristic approach the effectiveness of which we evaluate with experiments.

Comments

Permanent URL: http://dx.doi.org/10.7936/K7FT8JC4

Share

COinS