GTC 2020: Building MSOE's GPU-Powered Infrastructure for AI Instruction and Research
After clicking “Watch Now” you will be prompted to login or join.
Click “Watch Now” to login or join the NVIDIA Developer Program.
Building MSOE's GPU-Powered Infrastructure for AI Instruction and Research
Bradley Palmer , NVIDIA | Derek Riley, Milwaukee School of Engineering
Last September, Milwaukee School of Engineering, with NVIDIA, deployed a new hybrid cluster to serve multiple AI and HPC computing demands including instruction, student and faculty research, and industry collaborations. We'll describe the optimized accelerated computing architecture of this cluster and the software stack enabling these uses. We'll explain the lessons learned through the design, build, and deploy processes. We'll provide a blueprint for other institutions that seek to build GPU-accelerated computing infrastructure that supports similar diverse requirements.