The Parallel Computing Summer Research Internship is an intense 10-week program aimed at providing students with a solid foundation in modern high performance computing (HPC) topics integrated with research on real problems encountered in large-scale scientific codes.
Applications for 2026 are closed
During the 10-week program, students will receive training and lectures on modern topics in HPC and software development, including:
Students will collaborate in teams to identify and investigate different computational problems within the scientific focus area, and implement solutions guided by mentors with scientific and computational expertise. By working on cutting-edge HPC hardware, students will gain hands-on experience and learn how to effectively communicate their work through posters and oral presentations.
For the 2026 workshop, there will be 3 different projects. In the projects listed below, students will work in pairs and with their mentors who will oversee their project. The mentors are all established scientists at Los Alamos National Laboratory.
Nigel Tan, XCP-6, Ari Le, XCP-6, and Adam Stanier, T-5
Developing large scale multiphysics simulations requires many different physics components working together. Each component often has different time step sizes, mesh resolutions, performance properties, and scalability characteristics. You will work with physicists and HPC experts on porting an atmospheric air chemistry code from Python to C++ and parallelize it across different CPUs and GPUs using the Kokkos performance portability ecosystem. The air chemistry code will then be coupled with the Hybrid Vector Particle-In-Cell (VPIC) code as part of the Multiphysics Artificial Radiation Belt and E3 Electro-Magnetic Pulse High-Altitude Nuclear Explosion End-to-End Model (MAEHEM) project. You will gain valuable experience porting Python code to CPU and GPU based HPC platforms such as Rocinante, Chicoma, and Venado like systems. You will also learn about running multiphysics simulations on different HPC resources, using different profilers to identify performance bottlenecks, benchmarking code on a variety of different platforms, and how scientists turn research prototypes into practical production code.
Michael Witek, XCP-2 and Kyle Vaughn, XCP-2
Pagosa is a 3-D Eulerian hydrodynamics code that models phenomena associated with highvelocity impacts and explosively driven systems. Recent modernization efforts have enabled strong scaling to hundreds of nodes, but the code remains MPI-only. This limits its performance on hybrid CPU-GPU systems, e.g., LLNL’s El Capitan, which contains AMD MI300A APUs and has been verified as the world’s fastest supercomputer at 1.743 exaFLOPs, and LANL’s upcoming Mission and Vision systems that will be based on the NVIDIA Vera Rubin platform. With an MPI-only parallelization scheme, internal benchmarks have shown a performance bottleneck: as MPI ranks increase, they contend with a small set of GPUs, reducing overall efficiency. In this project, students will explore a hybrid MPI + OpenMP approach to solve the GPU contention problem. The students will develop benchmarks that demonstrate the feasibility of using a few MPI ranks to manage pools of OpenMP threads and communicate with GPUs. The students will then work to incorporate the scheme into the Pagosa codebase. Interested students should be proficient in modern C++ and will gain valuable experience with state-of-the-art HPC tools and modern software-engineering workflows for a large production code that runs on the world’s fastest supercomputers.
CJ Solomon (XCP-2) and Ivan Gonzalez (XCP-2)
Students will profile and improve the performance of the Eulerian Applications Project’s production code xRAGE. Students will profile and identify hotspots on the LANL HPC system Rocinante when using xRAGE’s on node OpenMP parallel computing capabilities. Depending on machine availability, they will also profile xRAGe on LANL’s Nvidia Grace+Hopper developmental cluster, Venadito. In addition to profiling experience, students will gain familiarity with C++ based performance portability layers, similar to the Linux Foundation’s Kokkos project, and develop an appreciation for the level of software engineering a production product such as xRAGE requires. This project will focus on improving the performance of on-GPU strength modelling in asteroid impact simulations.
Thomas Vogel (CAI-1) and Pete Maginot (XCP-2)
Students will profile and improve the performance of the Eulerian Applications Project’s production code xRAGE. Students will profile and identify hotspots on the LANL HPC system Rocinante when using xRAGE’s on node OpenMP parallel computing capabilities. Depending on machine availability, they will also profile xRAGe on LANL’s Nvidia Grace+Hopper developmental cluster, Venadito. In addition to profiling experience, students will gain familiarity with C++ based performance portability layers, similar to the Linux Foundation’s Kokkos project, and develop an appreciation for the level of software engineering a production product such as xRAGE requires. This project will focus on improving the performance of an ICF capsule simulation.