Profile Guided Dataflow Transformation for FPGAs and CPUs

STEWART, Robert, BHOWMIK, Deepayan, WALLACE, Andrew and MICHAELSON, Greg (2017). Profile Guided Dataflow Transformation for FPGAs and CPUs. Journal of Signal Processing Systems, 87 (1), 3-20.

[img]
Preview
PDF
jsps-2015.pdf - Accepted Version
Available under License All rights reserved.

Download (1MB) | Preview
Official URL: http://link.springer.com/article/10.1007/s11265-01...
Link to published version:: 10.1007/s11265-015-1044-y

Abstract

This paper proposes a new high-level approach for optimising field programmable gate array (FPGA) designs. FPGA designs are commonly implemented in low-level hardware description languages (HDLs), which lack the abstractions necessary for identifying opportunities for significant performance improvements. Using a computer vision case study, we show that modelling computation with dataflow abstractions enables substantial restructuring of FPGA designs before lowering to the HDL level, and also improve CPU performance. Using the CPU transformations, runtime is reduced by 43 %. Using the FPGA transformations, clock frequency is increased from 67MHz to 110MHz. Our results outperform commercial low-level HDL optimisations, showcasing dataflow program abstraction as an amenable computation model for highly effective FPGA optimisation.

Item Type: Article
Departments: Arts, Computing, Engineering and Sciences > Computing
Identification Number: 10.1007/s11265-015-1044-y
Depositing User: Deepayan Bhowmik
Date Deposited: 07 Nov 2016 14:24
Last Modified: 01 Aug 2017 13:02
URI: http://shura.shu.ac.uk/id/eprint/13807

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year

View more statistics