Accelerating and Scaling Python for HPC

Python is powering breakthrough exascale scientific discoveries - come learn how to program the world's largest supercomputers with it! In this interactive tutorial you’ll learn how to write, debug, profile, and optimize high-performance, multi-node GPU applications in Python.

You'll learn and master:

• CuPy for drop-in GPU acceleration of NumPy workflows,
• Nvmath-python for high level api for integrating Python with NVIDIA math libraries,
• Numba for writing custom kernels that match the performance of C++ and Fortran, and
• mpi4py for scaling across thousands of nodes.

Along the way we’ll learn how to profile our code, debug tricky kernels, and leverage foundational and domain-specific accelerated libraries.

Attendees should have a working knowledge of core Python syntax and some familiarity with NumPy array operations. No prior experience with GPUs, CUDA, or distributed computing is required, as these concepts are introduced from first principles during the hands-on sessions.

Participants are expected to attend in-person and actively participate in hands-on exercises. Remote attendance via zoom will be available, however, interaction with the instructors may be limited. Computing resources for interactive exercises may be limited to current NERSC users.