CHARM 2.0: Composing Heterogeneous Accelerators for Deep Learning on Versal ACAP Architecture
ACM Transactions on Reconfigurable Technology and Systems · 30.09.2024 · 17 (3) · 1-31
DOI: 10.1145/3686163 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM Transactions on Reconfigurable Technology and Systems, online erschienen am 30.09.2024. Verfasst von Jinming Zhuang und 13 weiteren Autor:innen. Die Arbeit wurde bisher 26-mal zitiert (Crossref) und nennt 54 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: NSF.
Abstract (Originalsprache)
Dense matrix multiply (MM) serves as one of the most heavily used kernels in deep learning applications. To cope with the high computation demands of these applications, heterogeneous architectures featuring both FPGA and dedicated ASIC accelerators have emerged as promising platforms. For example, the AMD/Xilinx Versal ACAP architecture combines general-purpose CPU cores and programmable logic with AI Engine processors optimized for AI/ML. An array of 400 AI Engine processors executing at 1 GHz can provide up to 6.4 TFLOPS performance for 32-bit floating-point (FP32) data. However, machine learning models often contain both large and small MM operations. While large MM operations can be parallelized efficiently across many cores, small MM operations typically cannot. We observe that executing some small MM layers from the BERT natural language processing model on a large, monolithic MM accelerator in Versal ACAP achieved less than 5% of the theoretical peak performance. Therefore, one key question arises: How can we design accelerators to fully use the abundant computation resources under limited communication bandwidth for end-to-end applications with multiple MM layers of diverse sizes? We identify the biggest system throughput bottleneck resulting from the mismatch between the massive computation resources of one monolithic accelerator and the various MM layers of small sizes in the application. To resolve this problem, we propose the CHARM framework to compose multiple diverse MM accelerator architectures working concurrently on different layers within one application. CHARM includes analytical models that guide design space exploration to determine accelerator partitions and layer scheduling. To facilitate system designs, CHARM automatically generates code, enabling thorough onboard design verification. We deploy the CHARM framework on four different deep learning applications in FP32, INT16, and INT8 data types, including BERT, ViT, NCF, and MLP, on the AMD/Xilinx Versal ACAP VCK190 evaluation board. Our experiments show that we achieve 1.46 TFLOPS, 1.61 TFLOPS, 1.74 TFLOPS, and 2.94 TFLOPS inference throughput for BERT, ViT, NCF, and MLP in FP32 data type, respectively, which obtain 5.29 \(\times\) , 32.51 \(\times\) , 1.00 \(\times\) , and 1.00 \(\times\) throughput gains compared to one monolithic accelerator. CHARM achieves the maximum throughput of 1.91 TOPS, 1.18 TOPS, 4.06 TOPS, and 5.81 TOPS in the INT16 data type for the four applications. The maximum throughput achieved by CHARM in the INT8 data type is 3.65 TOPS, 1.28 TOPS, 10.19 TOPS, and 21.58 TOPS, respectively. We have open-sourced our tools, including detailed step-by-step guides to reproduce all the results presented in this article and to enable other users to learn and leverage CHARM framework and tools in their end-to-end systems: https://github.com/arc-research-lab/CHARM .
Förderung
- NSF
Zitieren
APA
Zhuang, J., Lau, J., Ye, H., Yang, Z., Ji, S., Lo, J., Denolf, K., Neuendorffer, S., Jones, A., Hu, J., Shi, Y., Chen, D., Cong, J., & Zhou, P. (2024). CHARM 2.0: Composing Heterogeneous Accelerators for Deep Learning on Versal ACAP Architecture. ACM Transactions on Reconfigurable Technology and Systems, 17(3), 1-31. https://doi.org/10.1145/3686163
Vancouver
Zhuang J, Lau J, Ye H, Yang Z, Ji S, Lo J, et al. CHARM 2.0: Composing Heterogeneous Accelerators for Deep Learning on Versal ACAP Architecture. ACM Transactions on Reconfigurable Technology and Systems. 2024;17(3):1-31. doi:10.1145/3686163
BibTeX
@article{Zhuang2024,
author = {Zhuang, Jinming and Lau, Jason and Ye, Hanchen and Yang, Zhuoping and Ji, Shixin and Lo, Jack and Denolf, Kristof and Neuendorffer, Stephen and Jones, Alex and Hu, Jingtong and Shi, Yiyu and Chen, Deming and Cong, Jason and Zhou, Peipei},
title = {CHARM 2.0: Composing Heterogeneous Accelerators for Deep Learning on Versal ACAP Architecture},
journal = {ACM Transactions on Reconfigurable Technology and Systems},
year = {2024},
volume = {17},
number = {3},
pages = {1--31},
doi = {10.1145/3686163}
}
EUVIMED