VWR2A
Authors:
Open Access
- 10 July 2022
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 895-900
Abstract
Edge-computing requires high-performance energy-efficient embedded systems. Fixed-function or custom accelerators, such as FFT or FIR filter engines, are very efficient at implementing a particular functionality for a given set of constraints. However, they are inflexible when facing application-wide optimizations or functionality upgrades. Conversely, programmable cores offer higher flexibility, but often with a penalty in area, performance, and, above all, energy consumption. In this paper, we propose VWR2A, an architecture that integrates high computational density and low power memory structures (i.e., very-wide registers and scratchpad memories). VWR2A narrows the energy gap with similar or better performance on FFT kernels with respect to an FFT accelerator. Moreover, VWR2A flexibility allows to accelerate multiple kernels, resulting in significant energy savings at the application level.Keywords
Funding Information
- Botnar Foundation ReSoRT (REG-19-019)
- ERC Consolidator Grant COM- PUSAPIEN (725657)
- Swiss NSF ML-Edge (200020_182009)
This publication has 8 references indexed in Scilit:
- ULP-SRPACM Transactions on Reconfigurable Technology and Systems, 2014
- Energy-Efficient Communication Processors2013
- An Asymmetrical Register File: The VWR2010
- Very Wide Register: An Asymmetric Register File Organization for Low Power Embedded Processors2007
- Architecture Exploration for a Reconfigurable Architecture TemplateIEEE Design & Test of Computers, 2005
- Scratchpad memory2002
- MorphoSys: an integrated reconfigurable system for data-parallel and computation-intensive applicationsIEEE Transactions on Computers, 2000
- An algorithm for the machine calculation of complex Fourier seriesMathematics of Computation, 1965