Carry-Propagation-Adder-Factored Gemmini Systolic Array for Machine Learning Acceleration
Systolic arrays are the primary part of modern deep learning accelerators and are being used widely in real-life applications such as self-driving cars. This paper presents a novel factored systolic array, where the carry propagation adder for accumulation and the rounding logic are extracted out fr...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-03-01
|
Series: | Electronics |
Subjects: | |
Online Access: | https://www.mdpi.com/2079-9292/10/6/652 |
_version_ | 1797541738839015424 |
---|---|
author | Kashif Inayat Jaeyong Chung |
author_facet | Kashif Inayat Jaeyong Chung |
author_sort | Kashif Inayat |
collection | DOAJ |
description | Systolic arrays are the primary part of modern deep learning accelerators and are being used widely in real-life applications such as self-driving cars. This paper presents a novel factored systolic array, where the carry propagation adder for accumulation and the rounding logic are extracted out from each processing element, which reduces the area, power and delay of the processing elements substantially. The factoring is performed in the column-wise manner and the cost of the factored logic, placed at each column output, is amortized by the processing elements in a column. We demonstrate the proposed factoring in an open source systolic array, Gemmini. The factoring technique does not change the functionality of the base design and is transparent to applications. We show that the proposed technique leads to substantial reduction in area and delay up to 45.3% and 23.7%, respectively, compared to the Gemmini baseline. |
first_indexed | 2024-03-10T13:21:00Z |
format | Article |
id | doaj.art-1e6d91e98e944e9a9f958f40f097ac04 |
institution | Directory Open Access Journal |
issn | 2079-9292 |
language | English |
last_indexed | 2024-03-10T13:21:00Z |
publishDate | 2021-03-01 |
publisher | MDPI AG |
record_format | Article |
series | Electronics |
spelling | doaj.art-1e6d91e98e944e9a9f958f40f097ac042023-11-21T10:02:08ZengMDPI AGElectronics2079-92922021-03-0110665210.3390/electronics10060652Carry-Propagation-Adder-Factored Gemmini Systolic Array for Machine Learning AccelerationKashif Inayat0Jaeyong Chung1System on Chips Laboratory, Department of Electronics Engineering, Incheon National University, Incheon 22012, KoreaSystem on Chips Laboratory, Department of Electronics Engineering, Incheon National University, Incheon 22012, KoreaSystolic arrays are the primary part of modern deep learning accelerators and are being used widely in real-life applications such as self-driving cars. This paper presents a novel factored systolic array, where the carry propagation adder for accumulation and the rounding logic are extracted out from each processing element, which reduces the area, power and delay of the processing elements substantially. The factoring is performed in the column-wise manner and the cost of the factored logic, placed at each column output, is amortized by the processing elements in a column. We demonstrate the proposed factoring in an open source systolic array, Gemmini. The factoring technique does not change the functionality of the base design and is transparent to applications. We show that the proposed technique leads to substantial reduction in area and delay up to 45.3% and 23.7%, respectively, compared to the Gemmini baseline.https://www.mdpi.com/2079-9292/10/6/652machine learningGemminisystolic arrayfactorizationaccelerator |
spellingShingle | Kashif Inayat Jaeyong Chung Carry-Propagation-Adder-Factored Gemmini Systolic Array for Machine Learning Acceleration Electronics machine learning Gemmini systolic array factorization accelerator |
title | Carry-Propagation-Adder-Factored Gemmini Systolic Array for Machine Learning Acceleration |
title_full | Carry-Propagation-Adder-Factored Gemmini Systolic Array for Machine Learning Acceleration |
title_fullStr | Carry-Propagation-Adder-Factored Gemmini Systolic Array for Machine Learning Acceleration |
title_full_unstemmed | Carry-Propagation-Adder-Factored Gemmini Systolic Array for Machine Learning Acceleration |
title_short | Carry-Propagation-Adder-Factored Gemmini Systolic Array for Machine Learning Acceleration |
title_sort | carry propagation adder factored gemmini systolic array for machine learning acceleration |
topic | machine learning Gemmini systolic array factorization accelerator |
url | https://www.mdpi.com/2079-9292/10/6/652 |
work_keys_str_mv | AT kashifinayat carrypropagationadderfactoredgemminisystolicarrayformachinelearningacceleration AT jaeyongchung carrypropagationadderfactoredgemminisystolicarrayformachinelearningacceleration |