EDLAB : a benchmark for edge deep learning accelerators
A new trend tends to deploy deep learning algorithms to edge environments to mitigate privacy and latency issues from cloud computing. Diverse edge deep learning accelerators are devised to speed up the inference of deep learning algorithms on edge devices. Various edge deep learning accelerator...
Saved in:
Main Authors: | , , , , , , , , , , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | English |
Published: |
2022
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/155807 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
id |
sg-ntu-dr.10356-155807 |
---|---|
record_format |
dspace |
spelling |
sg-ntu-dr.10356-1558072023-12-15T03:01:03Z EDLAB : a benchmark for edge deep learning accelerators Kong, Hao Huai, Shuo Liu, Di Zhang, Lei Chen, Hui Zhu, Shien Li, Shiqing Liu, Weichen Rastogi, Manu Subramaniam, Ravi Athreya, Madhu Lewis, M. Anthony School of Computer Science and Engineering HP-NTU Digital Manufacturing Corporate Lab Engineering::Computer science and engineering Benchmark Deep Learning Edge Accelerator Deployment A new trend tends to deploy deep learning algorithms to edge environments to mitigate privacy and latency issues from cloud computing. Diverse edge deep learning accelerators are devised to speed up the inference of deep learning algorithms on edge devices. Various edge deep learning accelerators feature different characteristics in terms of power and performance, which make it a very challenging task to efficiently and uniformly compare different accelerators. In this paper, we introduce EDLAB, an end-to-end benchmark, to evaluate the overall performance of edge deep learning accelerators. EDLAB consists of state-of-the-art deep learning models, a unified workload preprocessing and deployment framework, as well as a collection of comprehensive metrics. In addition, we propose parameterized models to model the hardware performance bound so that EDLAB can identify the hardware potentials and the hardware utilization of different deep learning applications. Finally, we employ EDLAB to benchmark three edge deep learning accelerators and analyze the benchmarking results. From the analysis we obtain some insightful observations that can guide the design of efficient deep learning applications. Nanyang Technological University National Research Foundation (NRF) Submitted/Accepted version This research was conducted in collaboration with HP Inc. and supported by National Research Foundation (NRF) Singapore and the Singapore Government through the Industry Alignment Fund-Industry Collaboration Projects Grant (I1801E0028). This work is also partially supported by NTU NAP M4082282 and SUG M4082087, Singapore. 2022-03-22T02:21:08Z 2022-03-22T02:21:08Z 2021 Journal Article Kong, H., Huai, S., Liu, D., Zhang, L., Chen, H., Zhu, S., Li, S., Liu, W., Rastogi, M., Subramaniam, R., Athreya, M. & Lewis, M. A. (2021). EDLAB : a benchmark for edge deep learning accelerators. IEEE Design and Test. https://dx.doi.org/10.1109/MDAT.2021.3095215 2168-2356 https://hdl.handle.net/10356/155807 10.1109/MDAT.2021.3095215 en I1801E0028 NTU NAP M4082282 SUG M4082087 IEEE Design and Test 10.21979/N9/NT2HXS © 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/MDAT.2021.3095215. application/pdf |
institution |
Nanyang Technological University |
building |
NTU Library |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
NTU Library |
collection |
DR-NTU |
language |
English |
topic |
Engineering::Computer science and engineering Benchmark Deep Learning Edge Accelerator Deployment |
spellingShingle |
Engineering::Computer science and engineering Benchmark Deep Learning Edge Accelerator Deployment Kong, Hao Huai, Shuo Liu, Di Zhang, Lei Chen, Hui Zhu, Shien Li, Shiqing Liu, Weichen Rastogi, Manu Subramaniam, Ravi Athreya, Madhu Lewis, M. Anthony EDLAB : a benchmark for edge deep learning accelerators |
description |
A new trend tends to deploy deep learning algorithms to edge environments to mitigate privacy and latency issues from cloud computing.
Diverse edge deep learning accelerators are devised to speed up the inference of deep learning algorithms on edge devices.
Various edge deep learning accelerators feature different characteristics in terms of power and performance, which make it a very challenging task to efficiently and uniformly compare different accelerators.
In this paper, we introduce EDLAB, an end-to-end benchmark, to evaluate the overall performance of edge deep learning accelerators. EDLAB consists of state-of-the-art deep learning models, a unified workload preprocessing and deployment framework, as well as a collection of comprehensive metrics. In addition, we propose parameterized models to model the hardware performance bound so that EDLAB can identify the hardware potentials and the hardware utilization of different deep learning applications. Finally, we employ EDLAB to benchmark three edge deep learning accelerators and analyze the benchmarking results. From the analysis we obtain some insightful observations that can guide the design of efficient deep learning applications. |
author2 |
School of Computer Science and Engineering |
author_facet |
School of Computer Science and Engineering Kong, Hao Huai, Shuo Liu, Di Zhang, Lei Chen, Hui Zhu, Shien Li, Shiqing Liu, Weichen Rastogi, Manu Subramaniam, Ravi Athreya, Madhu Lewis, M. Anthony |
format |
Article |
author |
Kong, Hao Huai, Shuo Liu, Di Zhang, Lei Chen, Hui Zhu, Shien Li, Shiqing Liu, Weichen Rastogi, Manu Subramaniam, Ravi Athreya, Madhu Lewis, M. Anthony |
author_sort |
Kong, Hao |
title |
EDLAB : a benchmark for edge deep learning accelerators |
title_short |
EDLAB : a benchmark for edge deep learning accelerators |
title_full |
EDLAB : a benchmark for edge deep learning accelerators |
title_fullStr |
EDLAB : a benchmark for edge deep learning accelerators |
title_full_unstemmed |
EDLAB : a benchmark for edge deep learning accelerators |
title_sort |
edlab : a benchmark for edge deep learning accelerators |
publishDate |
2022 |
url |
https://hdl.handle.net/10356/155807 |
_version_ |
1787136805297979392 |