Enabling Graphics Processing Unit-based Data Science
UW–Madison researchers increasingly require graphics processing units (GPUs), specialized computing hardware that powers many of the dominant data science technologies, including but not limited to deep learning. This project will create GPU computing resources that include a community of technical experts, a library of sharable software and documentation, and a freely accessible pool of computer servers with GPUs managed by the Center for High Throughput Computing (CHTC) in order to unite existing GPU-based data science researchers and bring new tools into the hands of scientific domain experts.
These capabilities will support other emerging data science initiatives at the UW–Madison and accelerate and enable new research in a wide spectrum of areas including agronomy, bacteriology, botany, chemistry, computer science, computer vision, cryo-electron microscopy-based structural biology, engineering, genomics, machine learning, medical imaging, neuroscience, personalized medicine, physics and psychology.