Compiler-directed soft error resilience for lightweight GPU register file protection

Compiler-directed soft error resilience for lightweight GPU register file protection
复制标题

DOI:
10.1145/3385412.3386033
复制
发表时间:
2020-06
期刊:
Proceedings of the 41st ACM SIGPLAN Conference on Programming Language Design and Implementation
影响因子:
--
通讯作者:
Hongjune Kim;Jianping Zeng;Qingrui Liu;Mohammad Abdel-Majeed;Jaejin Lee;Changhee Jung
Hongjune Kim;Jianping Zeng;Qingrui Liu;Mohammad Abdel-Majeed;Jaejin Lee;Changhee Jung
中科院分区:
其他
文献类型:
--
作者:
Hongjune Kim;Jianping Zeng;Qingrui Liu;Mohammad Abdel-Majeed;Jaejin Lee;Changhee Jung

文献摘要

被引文献

相似文献

本文介绍了一分钱,这是一种以编译器指导的弹性计划,用于保护GPU寄存器文件(RF)免受软误差的范围,以替代常规的错误校正代码(ECC)的RF保护,使用较昂贵的错误检测代码(EDC)以及使用IDEMPOTS的恢复量相同,而Penny则可以实现同一级别的恢复。由于其能够检测到仅用于检测的多位数错误的能力,因此可以解决GPU中缺乏商店的缓冲区,这既导致检查点存储覆盖,penny的高成本,因此仅提供了几种编译器优化的储藏室,诸如存储着色和均等范围。
This paper presents Penny, a compiler-directed resilience scheme for protecting GPU register files (RF) against soft errors. Penny replaces the conventional error correction code (ECC) based RF protection by using less expensive error detection code (EDC) along with idempotence based recovery. Compared to the ECC protection, Penny can achieve either the same level of RF resilience yet with significantly lower hardware costs or stronger resilience using the same ECC due to its ability to detect multi-bit errors when it is used solely for detection. In particular, to address the lack of store buffers in GPUs, which causes both checkpoint storage overwriting and the high cost of checkpointing stores, Penny provides several compiler optimizations such as storage coloring and checkpoint pruning. Across 25 benchmarks, Penny causes only ≈3% run-time overhead on average.