Story of Two GPUs: Characterizing the Resilience of Hopper H100 and Ampere A100 GPUs
Authors:
- Shengkun CuiSCShengkun Cui
- Archit PatkeAPArchit Patke
- Hung NguyenHNHung Nguyen
- Aditya RanjanARAditya Ranjan
- Ziheng ChenZCZiheng Chen
- Phuong CaoPCPhuong Cao
- Gregory BauerGBGregory Bauer
- Brett BodeBBBrett Bode
- Catello Di MartinoCMCatello Di Martino
- Saurabh JhaSJSaurabh Jha
- Chandra NarayanaswamiCNChandra Narayanaswami
- Daby SowDSDaby Sow
Open Access
- 15 November 2025
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1145-1164
Abstract
No abstract availableAll Related Versions
Funding Information
- National Science Foundation (2029049,2319190,2430244,2530738,2320345,2005572,2138259,2138286,2138307,2137603,2138296)
- VinUni-Illinois Smart Health Center ()
- IBM-Illinois Discovery Accelerator Institute ()
This publication has 40 references indexed in Scilit:
- Advances in Appfl: a Comprehensive and Extensible Federated Learning Framework2025
- GPU Reliability Assessment: Insights Across the Abstraction Layers2024
- Software-controlled pipeline parity in GPU architectures for error detectionMicroelectronics Reliability, 2023
- G-SEAP: Analyzing and characterizing soft-error aware approximation in GPGPUsFuture Generation Computer Systems, 2020
- Impact of GPUs Parallelism Management on Safety-Critical and HPC Applications Reliability2014
- GPU Behavior on a Large HPC ClusterLecture Notes in Computer Science, 2014
- Hard Data on Soft Errors: A Large-Scale Assessment of Real-World Error Rates in GPGPU2010
- Scalable molecular dynamics with NAMDJournal of Computational Chemistry, 2005
- Modeling Coordinated Checkpointing for Large-Scale Supercomputers2005
- SLURM: Simple Linux Utility for Resource Management2003