ILP-Based and Heuristic Scheduling Techniques for Variable-Cycle Approximate Functional Units in High-Level Synthesis
Approximate computing is a promising approach to the design of area–power-performance-efficient circuits for computation error-tolerant applications such as image processing and machine learning. Approximate functional units, such as approximate adders and approximate multipliers, have been actively...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-09-01
|
Series: | Computers |
Subjects: | |
Online Access: | https://www.mdpi.com/2073-431X/11/10/146 |
_version_ | 1797474066088591360 |
---|---|
author | Koyu Ohata Hiroki Nishikawa Xiangbo Kong Hiroyuki Tomiyama |
author_facet | Koyu Ohata Hiroki Nishikawa Xiangbo Kong Hiroyuki Tomiyama |
author_sort | Koyu Ohata |
collection | DOAJ |
description | Approximate computing is a promising approach to the design of area–power-performance-efficient circuits for computation error-tolerant applications such as image processing and machine learning. Approximate functional units, such as approximate adders and approximate multipliers, have been actively studied for the past decade, and some of these approximate functional units can dynamically change the degree of computation accuracy. The greater their computational inaccuracy, the faster they are. This study examined the high-level synthesis of approximate circuits that take advantage of such accuracy-controllable functional units. Scheduling methods based on integer linear programming (ILP) and list scheduling were proposed. Under resource and time constraints, the proposed method tries to minimize the computation error of the output value by selectively multi-cycling operations. Operations that have a large impact on the output accuracy are multi-cycled to perform exact computing, whereas operations with a small impact on the accuracy are assigned a single cycle for approximate computing. In the experiments, we explored the trade-off between performance, hardware cost, and accuracy to demonstrate the effectiveness of this work. |
first_indexed | 2024-03-09T20:25:52Z |
format | Article |
id | doaj.art-1223c0fea1e741e48f18ff6d9342b009 |
institution | Directory Open Access Journal |
issn | 2073-431X |
language | English |
last_indexed | 2024-03-09T20:25:52Z |
publishDate | 2022-09-01 |
publisher | MDPI AG |
record_format | Article |
series | Computers |
spelling | doaj.art-1223c0fea1e741e48f18ff6d9342b0092023-11-23T23:35:59ZengMDPI AGComputers2073-431X2022-09-01111014610.3390/computers11100146ILP-Based and Heuristic Scheduling Techniques for Variable-Cycle Approximate Functional Units in High-Level SynthesisKoyu Ohata0Hiroki Nishikawa1Xiangbo Kong2Hiroyuki Tomiyama3Graduate School of Science and Engineering, Ritsumeikan University, Kusatsu 525-8577, Shiga, JapanGraduate School of Information Science and Technology, Osaka University, Suita 565-0871, Osaka, JapanGraduate School of Science and Engineering, Ritsumeikan University, Kusatsu 525-8577, Shiga, JapanGraduate School of Science and Engineering, Ritsumeikan University, Kusatsu 525-8577, Shiga, JapanApproximate computing is a promising approach to the design of area–power-performance-efficient circuits for computation error-tolerant applications such as image processing and machine learning. Approximate functional units, such as approximate adders and approximate multipliers, have been actively studied for the past decade, and some of these approximate functional units can dynamically change the degree of computation accuracy. The greater their computational inaccuracy, the faster they are. This study examined the high-level synthesis of approximate circuits that take advantage of such accuracy-controllable functional units. Scheduling methods based on integer linear programming (ILP) and list scheduling were proposed. Under resource and time constraints, the proposed method tries to minimize the computation error of the output value by selectively multi-cycling operations. Operations that have a large impact on the output accuracy are multi-cycled to perform exact computing, whereas operations with a small impact on the accuracy are assigned a single cycle for approximate computing. In the experiments, we explored the trade-off between performance, hardware cost, and accuracy to demonstrate the effectiveness of this work.https://www.mdpi.com/2073-431X/11/10/146high-level synthesisapproximate computingschedulingapproximate multiplication |
spellingShingle | Koyu Ohata Hiroki Nishikawa Xiangbo Kong Hiroyuki Tomiyama ILP-Based and Heuristic Scheduling Techniques for Variable-Cycle Approximate Functional Units in High-Level Synthesis Computers high-level synthesis approximate computing scheduling approximate multiplication |
title | ILP-Based and Heuristic Scheduling Techniques for Variable-Cycle Approximate Functional Units in High-Level Synthesis |
title_full | ILP-Based and Heuristic Scheduling Techniques for Variable-Cycle Approximate Functional Units in High-Level Synthesis |
title_fullStr | ILP-Based and Heuristic Scheduling Techniques for Variable-Cycle Approximate Functional Units in High-Level Synthesis |
title_full_unstemmed | ILP-Based and Heuristic Scheduling Techniques for Variable-Cycle Approximate Functional Units in High-Level Synthesis |
title_short | ILP-Based and Heuristic Scheduling Techniques for Variable-Cycle Approximate Functional Units in High-Level Synthesis |
title_sort | ilp based and heuristic scheduling techniques for variable cycle approximate functional units in high level synthesis |
topic | high-level synthesis approximate computing scheduling approximate multiplication |
url | https://www.mdpi.com/2073-431X/11/10/146 |
work_keys_str_mv | AT koyuohata ilpbasedandheuristicschedulingtechniquesforvariablecycleapproximatefunctionalunitsinhighlevelsynthesis AT hirokinishikawa ilpbasedandheuristicschedulingtechniquesforvariablecycleapproximatefunctionalunitsinhighlevelsynthesis AT xiangbokong ilpbasedandheuristicschedulingtechniquesforvariablecycleapproximatefunctionalunitsinhighlevelsynthesis AT hiroyukitomiyama ilpbasedandheuristicschedulingtechniquesforvariablecycleapproximatefunctionalunitsinhighlevelsynthesis |