Thesis: Optimized Code Generation of Parallel and Polyhedral Loop Nests using MLIR

PhD defense: Arun THANGAMANI

Team: ICPS

Date & time: Wednesday, September 25th, 2024, at 10:00 AM.
Location: Amphi A302, Pôle API, Blvd S. Brant, 67400 Illkirch-Graffenstaden.

Title: Optimized Code Generation of Parallel and Polyhedral Loop Nests using MLIR

Abstract: Our work shows the benefits of the novel MLIR compiler technology to the generation of code from a DSL, namely EasyML used in openCARP, a widely used simulator in the cardiac electrophysiology community. Building on an existing  work we deeply modified openCARP's native code generator to enable efficient vectorized  CPU  and GPU code generation (Nvidia CUDA and AMD ROCm). Generating optimized code for different accelerators requires specific optimizations and we review how MLIR has been used to enable multi-target code  generation from an integrated generator.
To our knowledge, this is the first work that deeply connects an optimizing compiler infrastructure to electrophysiology models of the human body, showing the potential benefits of using compiler technology in the simulation of human cell interactions.
Additionally, we did a study on the polyhedral compilers and generalized our techniques using Polygeist to improve the vectorization and heterogeneous code generation of polyhedral compilers.

The jury will be composed of:

• Prof. Stéphane Genaud, University of Strasbourg (thesis advisor)
• Dr. Vincent Loechner, University of Strasbourg (co-advisor)
• Dr. Guillaume Latu, Research Director, CEA Cadarache (reviewer)
• Prof. Alexandra Jimborean, University of Murcia (reviewer)
• Prof. Sylvain Contassot-Vivier, University of Lorraine (examiner)
• Prof. Mark Potse, University of Bordeaux (examiner)