
GRAY SCOTT SCHOOL 2025
The LAPP (CNRS/USMB), in collaboration with the CC-FR Competence Centre, is organizing the third Gray Scott School from 23 June to 4 July 2025. This summer school on HPC, in a unique format and entirely free of charge, will be dedicated to programming and optimization on Heterogeneous Architectures.
The school will cover the optimisation of computations on different types of hardware (CPU, GPU), presenting their respective characteristics, architectures and bottlenecks. It will cover generic optimisation methods applicable to all types of hardware, as well as the various libraries, technologies and languages available to achieve the best possible performance. Ideally, the peak performance of the machine.
- Hardware: CPU, GPU
- Languages considered : C++17, C++20, CUDA, Fortran, Rust, Python, Julia
- Libraries considered: SYCL, Eve, Numpy, cunumerics, legate, Jax, Thrust, cuPy, pycuda and PyTorch
- Compilers considered : G++, Clang++, nvc++, gfortran, nvfortran, dpc++. Profiling tools: Valgrind, Maqao, Perf, NSight, Malt and NumaProf
All the methods will be illustrated on simple examples, such as Hadamard products, reductions, barycentre calculations and matrix products, in order to be applied to a single problem: the simulation of a Gray Scott reaction.
This problem is simple enough to be understood quickly and complex enough for compilers to have difficulty optimising it without help. Each method will be broken down into a simple version, using default options, and one or more advanced versions, which will allow their advantages and disadvantages to be discussed and quantified.