NHR PerfLab Seminar: LLVM in HPC – Enabling Performance Portability, Interoperability, and Novel Features
Speaker: Dr. Johannes Doerfert, Lawrence Livermore National Laboratory
Title: LLVM in HPC – Enabling Performance Portability, Interoperability, and Novel Features
Date and time: Tuesday, May 30, 2:00 p.m. – 3:00 p.m. CET
Abstract:
In this talk we present recent efforts in and around the LLVM compiler framework targeting HPC codes with respect to performance, tooling, and development.
We will talk about portable CUDA, debugging and tuning at scale, effortless GPU execution of legacy codes, automatic differentiation of HPC programming models, ML in compilers, and the impact of missing static information.
While all our works are publicly available, we especially try to integrate them in the LLVM community compiler.
Since LLVM is often used directly, e.g., via LLVM/Clang, indirectly, e.g., via vendor compilers (AMD, Intel, HPE, …), or as part of other toolchains, e.g., Rust, Julia, and numba, we believe our R&D approach has the potential to profoundly impact HPC users and developers alike.
Short bio:
Johannes Doerfert is a computer scientist in the Center for Applied Scientific Computing at Lawrence Livermore National Laboratory. He is interested in new and exciting uses for compiler technologies. His research goal is to help people exploit hardware to the fullest without requiring them to become experts in the hardware or the software stack, including programming languages. Code is a means, not the final goal. As such, Johannes Doerfert believes that manual efforts to rewrite, tune, or adapt code are often signs for missing tools, compiler shortcomings, misinformation, or a combination thereof.
Johannes Doerfert is involved in the LLVM compiler framework since 2014 and the OpenMP language standard since 2018. Since 2021 he is the LLVM/OpenMP offload code owner. Johannes Doerfert received his Ph.D. in Computer Science from Saarland University in Germany in 2018.
For a list of past and upcoming NHR PerfLab seminar events, see: https://hpc.fau.de/research/nhr-perflab-seminar-series/