Andreas Kurth, Samuel Riedel, Florian Zaruba, Torsten Hoefler, Luca Benini:
ATUNs: Modular and Scalable Support for Atomic Operations in a Shared Memory Multiprocessor
(In Proceedings of the 57th Annual Design Automation Conference, ACM, Jun. 2020) Best Paper Finalist (6/228)
Abstract
Atomic operations are crucial for most modern parallel and concurrent algorithms, which necessitates their optimized implementation in highly-scalable manycore processors. We propose a modular and efficient, open-source ATomic UNit (ATUN) architecture that can be placed flexibly at different levels of the memory hierarchy. ATUN demonstrates near-optimal linear scaling for various synthetic and real-world workloads on an FPGA prototype with 32 RISC-V cores. We characterize the hardware complexity of our ATUN design in 22 nm FDSOI and find that it scales linearly in area (only 0.5 kGE per core) and logarithmically in the critical path.
Documents
download article:
Recorded talk (best effort)
BibTeX
@inproceedings{kurth-atun, author={Andreas Kurth and Samuel Riedel and Florian Zaruba and Torsten Hoefler and Luca Benini}, title={{ATUNs: Modular and Scalable Support for Atomic Operations in a Shared Memory Multiprocessor}}, year={2020}, month={Jun.}, booktitle={Proceedings of the 57th Annual Design Automation Conference}, publisher={ACM}, source={http://www.unixer.de/~htor/publications/}, }
