Publications
Publications & talks
Selected academic work from my PhD in computer architecture at UT Austin — resilience, GPU-dense systems, and runtime support for fault tolerance.
Papers
GPU Snapshot: Checkpoint Offloading for GPU-Dense Systems
Kyushick Lee, Michael Sullivan, Siva Kumar Sastry Hari, Timothy Tsai, Stephen W. Keckler, Mattan Erez
Architectural support for "checkpoint offloading" in GPU-dense systems.
On the Trend of Resilience for GPU-Dense Systems
Kyushick Lee, Michael Sullivan, Siva Kumar Sastry Hari, Timothy Tsai, Stephen W. Keckler, Mattan Erez
Effect of GPU density on multi-level checkpointing systems at scale.
Containment Domains Semantics, version 0.2
Michael Sullivan, Ikhwan Lee, Jinsuk Chung, Kyushick Lee, Song Zhang, Seong-Lyong Gong, Derong Liu, Mattan Erez
Concepts and semantics of Containment Domains (CDs).
Resilience À la carte: Application Tailored Resilience in Legion
Karthik Murthy, Mike Bauer, Kyushick Lee, Alex Aiken, Mattan Erez, et al.
Application-tailored resilience in the Legion task-based runtime.
Talks
- Resilient Heterogeneous System with Containment Domains — SW Engineering Retreat, UT Austin, 2018
- Resilient Exascale System with Containment Domains — Google Accelerator Summit, 2017
- Demo of CD Runtime for MPI Programs — Demo, 2016
The complete legacy academic site is preserved at kyushick.github.io.