AHA: An Agile Approach to the Design of Coarse-Grained Reconfigurable Accelerators and Compilers
ACM Transactions on Embedded Computing Systems · 31.03.2023 · 22 (2) · 1-34
DOI: 10.1145/3534933 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM Transactions on Embedded Computing Systems, online erschienen am 24.01.2023. Verfasst von Kalhan Koul und 32 weiteren Autor:innen. Die Arbeit wurde bisher 42-mal zitiert (Crossref) und nennt 61 Literaturangaben. Gefördert von: DSSoC DARPA, Stanford AHA Agile Hardware Center, Affiliates Program, Intel’s Science and Technology Center.
Abstract (Originalsprache)
With the slowing of Moore’s law, computer architects have turned to domain-specific hardware specialization to continue improving the performance and efficiency of computing systems. However, specialization typically entails significant modifications to the software stack to properly leverage the updated hardware. The lack of a structured approach for updating the compiler and the accelerator in tandem has impeded many attempts to systematize this procedure. We propose a new approach to enable flexible and evolvable domain-specific hardware specialization based on coarse-grained reconfigurable arrays (CGRAs). Our agile methodology employs a combination of new programming languages and formal methods to automatically generate the accelerator hardware and its compiler from a single source of truth. This enables the creation of design-space exploration frameworks that automatically generate accelerator architectures that approach the efficiencies of hand-designed accelerators, with a significantly lower design effort for both hardware and compiler generation. Our current system accelerates dense linear algebra applications but is modular and can be extended to support other domains. Our methodology has the potential to significantly improve the productivity of hardware-software engineering teams and enable quicker customization and deployment of complex accelerator-rich computing systems.
Förderung
- DSSoC DARPA
- Stanford AHA Agile Hardware Center
- Affiliates Program, Intel’s Science and Technology Center
- Stanford SystemX Alliance
Zitieren
APA
Koul, K., Melchert, J., Sreedhar, K., Truong, L., Nyengele, G., Zhang, K., Liu, Q., Setter, J., Chen, P., Mei, Y., Strange, M., Daly, R., Donovick, C., Carsello, A., Kong, T., Feng, K., Huff, D., Nayak, A., Setaluri, R., …, & Raina, P. (2023). AHA: An Agile Approach to the Design of Coarse-Grained Reconfigurable Accelerators and Compilers. ACM Transactions on Embedded Computing Systems, 22(2), 1-34. https://doi.org/10.1145/3534933
Vancouver
Koul K, Melchert J, Sreedhar K, Truong L, Nyengele G, Zhang K, et al. AHA: An Agile Approach to the Design of Coarse-Grained Reconfigurable Accelerators and Compilers. ACM Transactions on Embedded Computing Systems. 2023;22(2):1-34. doi:10.1145/3534933
BibTeX
@article{Koul2023,
author = {Koul, Kalhan and Melchert, Jackson and Sreedhar, Kavya and Truong, Leonard and Nyengele, Gedeon and Zhang, Keyi and Liu, Qiaoyi and Setter, Jeff and Chen, Po-Han and Mei, Yuchen and Strange, Maxwell and Daly, Ross and Donovick, Caleb and Carsello, Alex and Kong, Taeyoung and Feng, Kathleen and Huff, Dillon and Nayak, Ankita and Setaluri, Rajsekhar and Thomas, James and Bhagdikar, Nikhil and Durst, David and Myers, Zachary and Tsiskaridze, Nestan and Richardson, Stephen and Bahr, Rick and Fatahalian, Kayvon and Hanrahan, Pat and Barrett, Clark and Horowitz, Mark and Torng, Christopher and Kjolstad, Fredrik and Raina, Priyanka},
title = {AHA: An Agile Approach to the Design of Coarse-Grained Reconfigurable Accelerators and Compilers},
journal = {ACM Transactions on Embedded Computing Systems},
year = {2023},
volume = {22},
number = {2},
pages = {1--34},
doi = {10.1145/3534933}
}
EUVIMED