2026-07-13 –, Accelerated Computing
Scientific researchers need reproducible software environments for complex applications that can run across heterogeneous computing platforms. Modern open source tools, like Pixi, provide automatic reproducibility solutions for all dependencies while providing a high level interface well suited for researchers.
This tutorial will provide a practical introduction to using Pixi to easily create scientific and AI/ML environments that benefit from hardware acceleration, across multiple machines and platforms. The focus will be on CUDA applications, such as machine learning frameworks and use of CUDA Tile, as well as using pixi-build to construct bespoke CUDA enabled conda packages.
Installation Instructions: https://matthewfeickert-talks.github.io/reproducible-cuda-workflows-with-pixi-scipy-2026/setup/
As artificial intelligence (AI) and machine learning (ML) becomes a modern part of the scientific toolkit, the need to have robustly reproducible scientific computing environments that support hardware acceleration, e.g. with CUDA, becomes more important. However, historically just installing a working CUDA environment on a single machine, let alone on multiple platforms with different requirements, could be a difficult task for non-experts. This led to many scientific machine learning workflows being reliably runnable on only particular machines, and, even worse, with environments that were not reproducible across time.
With significant recent advancements by the NVIDIA open source team and the conda-forge open source community, the entire CUDA stack — from compilers to runtime libraries — is now distributed on conda-forge. This significantly reduces the overhead to install CUDA dependencies, but packaging and distribution of binaries alone does not solve the problem of reproducibility. With automatic multi-platform hash-level lock file support for all dependencies that are available on package indexes (like PyPI and conda-forge), highly efficient solving strategies, and high level user interfaces, Pixi provides a missing piece to the scientific researcher toolkit. With Pixi, researchers are able to easily specify the hardware acceleration requirements they have, multiple different computational environments needed for their experiments, and the required software dependencies, and then quickly solve for a multi-platform lock file of all the dependencies required, down to the compiler level. This makes it possible to have multiple hardware accelerated environments defined that are able to run hardware accelerated workflows across heterogeneous machines with different GPU types and CUDA compatibility.
This tutorial will be targeted to scientific researchers who use Python for scientific computing and use hardware accelerated workflows in their research, with a particular focus on AI/ML. No prior expertise with hardware accelerator systems is assumed. The tutorial structure will begin with an introduction to Pixi as a computational environment manager, and explore how it provides features beyond other more common package managers that might be used for Python dependencies. It will then extend to adding CUDA requirements to Pixi environments, and provide participants with exercises for solving environments and running simple AI/ML workflows using the PyTorch machine learning library and the cuTile Python library. The tutorial will then move towards more complex environment requirements in later exercises. The tutorial will conclude with examples and exercises on building bespoke CUDA enabled conda packages with pixi-build.
Tutorial participants will code all examples themselves. Participants will also be given time to explore solutions to their own hardware accelerated Python workflows. To make the tutorial more practical and interactive, NVIDIA has agreed to donate cloud GPU resources on the NVIDIA Brev platform, which will allow for participants to have CUDA enabled GPU resources to run their own examples on.
Participants should be familiar with Python programming for science, and using external dependencies in their work. The tutorial will use machine learning workflows as examples, but while familiarly with machine learning may be useful for conceptual understanding of the tasks, no prior machine learning knowledge is required to complete the tutorial. No prior expertise with CUDA is assumed.
Installation Instructions:Pixi is the only tool that needs to be installed prior to the start of the tutorial. Install instructions for Pixi are provided on the Pixi documentation website, but can be summarized as * Linux, macOS: curl -fsSL https://pixi.sh/install.sh | bash * Windows powershell -ExecutionPolicy ByPass -c "irm -useb https://pixi.sh/install.ps1 | iex"
Matthew is a research scientist in experimental high energy physics and data science at the University of Wisconsin-Madison Data Science Institute (a “data physicist”). He works as a member of the ATLAS collaboration on searches for physics beyond the standard model with experiments performed at CERN's Large Hadron Collider (LHC) in Geneva, Switzerland. He also serves on the executive board of the Institute for Research and Innovation in Software for High Energy Physics (IRIS-HEP) where he is a researcher and the Analysis Systems Area lead. He is also a topical editor for physics and data science for the Journal of Open Source Software. He previously did his Ph.D. (2019) research at Southern Methodist University, also on the ATLAS experiment, and was a postdoc at the University of Illinois at Urbana-Champaign, and the University of Wisconsin-Madison.
Ruben is part of the Prefix.dev core team, builing Pixi and other tools in the package management space. Originally he's a Robotics engineer working on industrial robots, but quickly figuring out that solving development and deployment problems were one of the bigger issues that robotics developers had to deal with. Joining Prefix.dev allowed him to focus on improving the UX/DX of a large group of software engineers. Over the years he's been doing multiple talks and workshops on how to properly manage software and development workflows.
Dr. Katrina Riehl is a Principal Technical Product Manager at NVIDIA leading the CUDA Education program. For over two decades, Katrina has worked extensively in the fields of scientific computing, machine learning, data science, and visualization. Most notably, she has helped lead data initiatives at the University of Texas Austin Applied Research Laboratory, Anaconda, Apple, Expedia Group, Cloudflare, and Snowflake. She is an active volunteer in the Python open-source scientific software community and currently serves on the Advisory Council for NumFOCUS.