Leiden University
Company Website
Full Time
PhD Candidate, Energy Efficient GPU Computing

PhD Candidate, Energy Efficient GPU Computing

We are looking for a fully funded PhD candidate in High Performance Computing (HPC). Powerful computers are driving the rapid advancements in science, industry, and society, but at huge energy and environmental costs. Software is often not fully optimized to use the underlying hardware in the most optimal way, leaving room to improve the energy efficiency. In this project, we will focus on automatic performance tuning (auto-tuning) of applications for Graphics Processing Units (GPUs) for energy efficiency. This project will research new algorithms and methods to automatically optimize software for energy efficiency.

The specific focus of the research is to be determined in consultation with the supervisor, dr. Ben van Werkhoven. The research will be related to Kernel Tuner, a software development tool for automatic performance tuning (auto-tuning) GPU applications. Kernel Tuner is developed by the Accelerated Computing group and several national and international collaborators.