The Accelerator Store framework for high-performance, low-power accelerator-based systems

Authors:
Michael Lyons;Mark Hempstead;Gu-Yeon Wei;David Brooks
Affiliations:
Harvard University, Cambridge;Drexel University, Philadelphia;Harvard University, Cambridge;Harvard University, Cambridge
Venue:
IEEE Computer Architecture Letters
Year:
2010

Citing 0
Cited 3

The accelerator store: A shared memory framework for accelerator-based systems

ACM Transactions on Architecture and Code Optimization (TACO) - HIPEAC Papers
Progress on 'pico' air vehicles

International Journal of Robotics Research
Low complexity QR-decomposition architecture using the logarithmic number system

Proceedings of the Conference on Design, Automation and Test in Europe

Quantified Score

Hi-index	0.00

Visualization

Abstract

Hardware acceleration can increase performance and reduce energy consumption. To maximize these benefits, accelerator- based systems that emphasize computation on accelerators (rather than on general purpose cores) should be used. We introduce the “accelerator store,” a structure for sharing memory between accelerators in these accelerator-based systems. The accelerator store simplifies accelerator I/O and reduces area by mapping memory to accelerators when needed at runtime. Preliminary results demonstrate a 30% system area reduction with no energy overhead and less than 1% performance overhead in contrast to conventional DMA schemes.