workshoprexio252025

Optimizing I/O for an Exascale Implicit Kinetic Plasma Simulation using the Rabbit Storage System

Ian Lumsden, Hariharan Devarajan, Izzet Yildirim, Stefano Markidis, Amy Hu, Ivy Peng, Lorenzo Pennati, David Yokelson, Stephanie Brink, Olga Pearce, Tom Scogland, Bronis R. Supinski, Gian Luca Delzanno, Anthony Kougkas, Xian-He Sun, Michela Taufer

REX-IO 2025 in conjunction with IEEE Cluster 2025 · IEEE

DOI ↗

Abstract

Exascale systems present unprecedented I/O challenges due to their massive scale and complex memory hierarchies. This work presents optimizations for an implicit kinetic plasma simulation using the Rabbit storage system and studies I/O accelerators and burst buffers on exascale platforms.

Connected work

Iowarp

Citation

@inproceedings{lumsden2025rabbit,
  author = {Lumsden, Ian and Devarajan, Hariharan and Yildirim, Izzet and Markidis, Stefano and Hu, Amy and Peng, Ivy and Pennati, Lorenzo and Yokelson, David and Brink, Stephanie and Pearce, Olga and Scogland, Tom and de Supinski, Bronis R. and Delzanno, Gian Luca and Kougkas, Anthony and Sun, Xian-He and Taufer, Michela},
  booktitle = {REX-IO 2025 in conjunction with IEEE Cluster 2025},
  title = {Optimizing I/O for an Exascale Implicit Kinetic Plasma Simulation using the Rabbit Storage System},
  year = {2025},
  month = sep,
  publisher = {IEEE},
  volume = {},
  number = {},
  pages = {},
  keywords = {HPC, Rabbit Nodes, I/O Accelerators, I/O Optimization, iPIC3D, IOR},
  doi = {10.1109/CLUSTERWorkshops65972.2025.11164212},
  url = {https://doi.org/10.1109/CLUSTERWorkshops65972.2025.11164212},
}