Stanford University logo SLAC National Accelerator Laboratory logo Los Alamos National Laboratory logo NVIDIA logo Winner of the R&D 100 Award

Legion

A Data-Centric Parallel Programming System

Github

Legion Retreat 2024

Session 1: Project Updates

  1. Opening Remarks – Alex Aiken, Stanford [video]
  2. Legion – Michael Bauer, NVIDIA [video]
  3. Regent & Pygion – Elliott Slaughter, SLAC [video]
  4. Realm – Artem Priakhin, NVIDIA [video]

Session 2: Legate

  1. Legate and cuPyNumeric – Wonchan Lee, NVIDIA [video]
  2. Legate Jax – Jeremy Wilke, NVIDIA [video]
  3. Legate Sparse – Rohan Yadav, Stanford [video]

Session 3: Applications I

  1. Ristra, FlecSI, and Legion: Multi-Physics on Unstructured Meshes – Jonathan Graham, Los Alamos National Laboratory [video]
  2. Real-time Aperture Synthesis Imaging with the DSA-2000 Radio Camera – Martin Pokorny, CalTech [video]
  3. Accelerated Massive Data Analytics for X-Ray Imaging of Materials and Semiconductors – Quynh Nguyen, SLAC [video]

Session 4: Short Talks and Discussion

  1. Are Task-Based Programming Models Suitable for Processing in Memory (PIM) Clusters? – David Krasowska, Northwestern University [video]
  2. Regent-FFT: A Fast Fourier Transform Library for Regent – Arjun Kunna, Stanford [video]

Session 5: Applications II

  1. HTR Solver: Studying Compressible Reacting Flows Using Legion – Mario Di Renzo, University of Salento [video]
  2. Low-Latency, High-Performance LLM Serving and Fine-tuning with Legion/FlexFlow – Zhihao Jia, Carnegie Mellon University [video]
  3. FlexFlow – Colin Unger, Stanford University [video]

Session 6: Debugging

  1. The Legion Profiler & Fuzzer – Elliott Slaughter, SLAC [video]
  2. Debugging Legion Applications – Michael Bauer, NVIDIA [video]