GPU Performance Modeling

Berkeley Lab is a world leader in computing, biological, physical, materials, chemical, and energy sciences research. As a summer student, you will work on a project within one of the four divisions in the Computing Sciences Area: the National Energy Research Scientific Computing Center (NERSC). NERSC is the primary scientific computing facility for the Office of Science in the U.S. Department of Energy. As one of the largest facilities in the world devoted to providing computational resources and expertise for basic scientific research, NERSC is a world leader in accelerating scientific discovery through computation.

If you are a student majoring in computational science, computer science, data science, mathematics, or related fields, join NERSC for a 10–12 week research internship to gain experience in computational research and high performance computing under the guidance of one or more staff members.

What you will do

In this exciting internship, you will study the performance of math libraries on CPUs and GPUs through performance modeling tools. You will develop interactive, browser-based labs that teach users how to model application performance on NERSC’s GPU-accelerated systems. The specific tasks involved include the following:

• Develop a Jupyter notebook with interactive elements that run on NERSC systems.
• Create a “GPU Performance Modeling” course module on Canvas with embedded code execution and performance visualization tools.
• Model the performance of a math library on GPUs.

Requirements

Applicants must be enrolled in a full-time academic program at an accredited college or university and majoring in computational science, computer science, data science, mathematics, or a related field. Proof of enrollment is required.

Desired skills/background

• Experience with Jupyter
• Experience with parallel codes and/or GPU codes
• Experience with performance modeling

Back to 2026 projects