PACE: Power-Aware Computing Engines

Abstract

This report describes the PACE project whose objective was to reduce the energy consumption of microprocessors by exploiting compile time knowledge to reduce run-time switching activity and to power down unneeded blocks. The project had two phases. The first phase focused on understanding and reducing power consumption within microprocessor components, such as caches, register files, and arithmetic units. Several new techniques were developed to reduce both switching and leakage power. The second phase developed a new energy-exposed microprocessor architecture, SCALE (Software-Controlled Architecture for Low Energy). SCALE is based on a new vector-thread architectural paradigm which unifies the vector and threaded execution models, to provide efficient execution of many forms of parallelism. The SCALE vector thread architecture and the detailed design are being pursued in other projects. The PACE project developed a variety of power saving techniques at both the micro architectural and instruction set level, several of which are being actively transferred to industry. Over a dozen conference papers and student theses have been published to distribute results to the research community.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 01, 2005
Accession Number
ADA431169

Entities

People

  • Krste Asanovic

Organizations

  • Massachusetts Institute of Technology

Tags

Communities of Interest

  • Air Platforms
  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Air Force Research Laboratories
  • Artificial Intelligence
  • Coding
  • Computer Architecture
  • Computer Programming
  • Computer Programs
  • Computer Science
  • Computers
  • Computing System Architectures
  • Data Compression
  • Embedded Systems
  • Energy Consumption
  • Instruction Set Architecture
  • Network Science
  • Operating Systems
  • Trees (Data Structures)
  • Wireless Communications

Fields of Study

  • Computer science
  • Engineering

Readers

  • Defense Technology Research and Development.
  • Parallel and Distributed Computing.