Students at Southern Methodist University in Dallas, Texas, put together 16 Nvidia Jetson Nanos to create a “baby supercomputer.” (opens in new tab) It comes with a module, a network switch, some power supplies, and a cooling fan. The resulting cluster is Nvdia Blog, (opens in new tab) It is a size that can be placed on a desk and can be used as an educational tool.
Built from scratch in 4 months by 2 students after receiving a small grant to create a Jetson cluster. The student also handcrafted an acrylic case, over 60 interconnecting cables, and a power supply to learn about her wire stripping. Networking on the way, and a laser cutter. (There was only one small fire.) The computer was born as a collection of Jetson Nano modules sprawled out on your desk, using cardboard boxes as “heat sinks,” but the 3D vector files that feed the lasers. was upgraded to the current neat acrylic case. A cutter in the university’s makerspace.
Eric Godat, Team Leader for Research and Data Science in SMU’s internal IT organization, said: “Our Nvidia DGX SuperPOD just opened on campus, so we don’t need this baby supercomputer to be a real computing environment.” You can try reimaging the , and deploying the cluster software. ”
Raspberry Pi for AI (opens in new tab)Jetson Nano features a 64-bit quad-core Arm CPU and a 128-core Nvidia Maxwell GPU, with 4 GB of RAM and USB, Ethernet, HDMI and M.2 sockets. It also has a 40-pin GPIO array and typically runs a version of Ubuntu specifically tuned for the hardware. The cluster’s software environment is still a work in progress and is built using Nvidia Jetpack and will be used for small machine learning tasks.
The Jetson cluster is SC22 Supercomputing Conference (opens in new tab) This year it will be held in Dallas from November 13-18.