Tiny Titan Preps Students for Multicore Era
Supercomputing can be daunting to the uninitiated. A peek at the hardware, for example, with row after row of black boxes, reveals little of the exciting science that is being enabled by these ultra-fast computations. To help make the world of HPC a little less opaque, staff members at Oak Ridge National Labs (ORNL) created a scaled down version of the Titan supercomputer. The team developed a nine-core portable unit that communicates its core activity with lights.
This scaled down supercomputer looks a bit like a child’s toy, but there’s a reason for the multi-colored madness. Each color represents a processor and images on the connected monitor use the same colors to show what each processor is doing. The more colors that light up, the faster the program is running. The fewer the colors, the slower the execution.
“It’s a lot better when you can actually visualize the difference,” says Austin Peay State University student Samuel Cupp.
Program backers are using the tiny supercomputer to introduce students to parallel computing. It’s the future of computer science education, says Robert French of the OLCF User Assistance and Outreach group. “Everything from your smart phone to your 3D television has multicore processors, so students that don’t have parallel programming skills will be left out of 21st century engineering jobs,” he adds.
French presented Tiny Titan at an ORNL event called the Next Big Idea Competition. While it did not win a monetary award, it did receive the People’s Choice Award.
The problem with current HPC outreach is that many programs are still teaching serial processing or computer processing on just one core.
“Students need to learn to program in parallel,” maintains French. “Computers will continue to get more cores and become more complicated.”
The scaled down systems were constructed using Raspberry Pi Foundation ARM-based microcomputers, at a cost of just $35 each. The Pi chips, like other ultra-low-power processors on the market, are economical and can be combined with other components to build an affordable multicore processor. The details are shared here.
Now that they’ve built the system, the next step for the Tiny Titan development team is to create a curriculum for teaching parallel computing in high schools and STEM programs. The group is working to get the system in front of students at local schools. Their initial outreach efforts will focus on schools that already have STEM programs in place.
Tiny Titan was recently featured on a local news show, which produced the clip below.