Bayesian Optimization for Efficient Accelerator Synthesis
- 30 December 2020
- journal article
- research article
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Architecture and Code Optimization
- Vol. 18 (1), 1-25
- https://doi.org/10.1145/3427377
Abstract
Accelerator design is expensive due to the effort required to understand an algorithm and optimize the design. Architects have embraced two technologies to reduce costs. High-level synthesis automatically generates hardware from code. Reconfigurable fabrics instantiate accelerators while avoiding fabrication costs for custom circuits. We further reduce design effort with statistical learning. We build an automated framework, called Prospector, that uses Bayesian techniques to optimize synthesis directives, reducing execution latency and resource usage in field-programmable gate arrays. We show in a certain amount of time that designs discovered by Prospector are closer to Pareto-efficient designs compared to prior approaches. Prospector permits new studies for heterogeneous accelerators.Keywords
This publication has 24 references indexed in Scilit:
- A Survey and Evaluation of FPGA High-Level Synthesis ToolsIEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2015
- Taking the Human Out of the Loop: A Review of Bayesian OptimizationProceedings of the IEEE, 2015
- CAPI: A Coherent Accelerator Processor InterfaceIBM Journal of Research and Development, 2015
- SPIRIT: Spectral-Aware Pareto Iterative Refinement Optimization for Supervised High-Level SynthesisIEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2014
- Machine learning predictive modelling high-level synthesis design space explorationIET Computers & Digital Techniques, 2012
- The gem5 simulatorACM SIGARCH Computer Architecture News, 2011
- Efficient design space exploration for application specific systems-on-a-chipJournal of Systems Architecture, 2007
- Cache optimization for embedded processor coresACM Transactions on Design Automation of Electronic Systems, 2004
- The balance between proximity and diversity in multiobjective evolutionary algorithmsIEEE Transactions on Evolutionary Computation, 2003
- A fast and elitist multiobjective genetic algorithm: NSGA-IIIEEE Transactions on Evolutionary Computation, 2002