PRUNER

Providing Reproducibility on Ubiquitously Non-deterministic Environments and Runs


Description

This is a project launched by Lawrence Livermore National Laboratory (LLNL) in collaboration with the Formal Verification Group at School of Computing of University of Utah. This project seeks to identify and solve some of the variability-related problems that, if left to chance or ad hoc solutions, can impede continued progress in the field of Extreme Scale computing.
In detail, the project consists in building a tool for helping programmers in identification of non-determinism results and data races in parallel programs that involve paradigms of concurrency in high-performance computing, namely OpenMP and MPI.

Experiments

Publications

  • Determinism and Reproducibility in Large-Scale HPC Systems - Wei-Fan Chiang, Ganesh Gopalakrishnan, Zvonimir Rakamaric, Dong H. Ahn and Gregory L. Lee [paper]
  • Overcoming Extreme-Scale Reproducibility Challenges Through a Unified, Targeted, and Multilevel Toolset - Dong H. Ahn, Gregory L. Lee, Ganesh Gopalakrishnan, Zvonimir Rakamaric, Martin Schulz and Ignacio Laguna [paper]

Project Members

  • Professors: Ganesh Gopalakrishnan, Zvonimir Rakamaric
  • PhD Students: Simone Atzeni
  • LLNL: Dong H. Ahn, Chris Chambreau, Gregory L. Lee, Ignacio Laguna, Martin Schulz