Consumer video game hardware proves useful for science

By Sebastian Moronta Blanco, Staff Writer

The hardware developed for video gaming such as consoles, tablets, and the graphics processing parts contained within can be compared closely to leading supercomputer hardware, according to researcher Glenn Volkema.

On November 8, Glenn Volkema, a member of the Physics Department and the Center for Scientific Computing and Visualization Research at UMass Dartmouth, gave a presentation on the above concept to computer science and physics majors in the Textile Building.

Volkema and his colleagues have been researching the use of video gaming hardware for computing purposes for some time now, co-publishing an article in July with Dr. Gaurav Khanna, a physics professor at UMass Dartmouth, detailing the tests and results that were shared in the presentation this past Tuesday.

Dr. Khanna and his team also conducted a project in 2007 to use multiple PlayStation 3’s connected together to combine their processing power into a single computer. The cluster used eight PS3’s and the team published scientific results, eventually leading the U.S. Air Force Research Laboratory to build the “Condor Cluster”, a supercomputer composed of 1,760 PS3’s that was at the time recognized as the 33rd largest supercomputer in the world.

Volkema presented the results of his team’s series of tests conducted on hardware designed for video gaming as compared to their top of the line counter-parts, and the tests compared hardware from two different tech companies: NVidia and AMD. Both companies offer several different lines of graphics processors, and the tests took the NVidia Tegra X1 and the AMD Radeon R9 Fury X, two GPU’s meant for video gaming, and compared them to the NVidia Tesla K40 and the AMD A10-7850K, GPU’s sold for supercomputers.

The team also included the AMD FX-9590, a traditional processor that was included solely for the purposes of comparison.

The tests involved using the OpenCL SHOC benchmark suite, which uses multiple measures to test computing power, and comparing results. The team also used Einstein@Home, a publicly distributed program that uses computers to search for gravitational waves in outer space, to test the GPU’s processing power as well.

The results showed gaming GPU’s performing at just below the advanced GPU’s while operating on a fraction of the power and at a fraction of the cost. The top of the line models can cost several thousand dollars, however, the GPU’s sold to traditional consumers for gaming typically cost under one thousand.

Volkema, in an interview with Next Platform, explained the results as “low cost and high power efficiency. Rapid advances and significant innovation is being enabled through major investments made by the gaming industry. This is driven by strong consumer demand for immersive gaming experiences that require computational power. High volume and intense competition keep costs low, while improvements to power efficiency are forced by the engineering challenges and costs associated with dissipating increasing amounts of waste heat from a discrete device and the limited battery life of a mobile device.”

In other words, competition and high demand have forced gaming tech companies to produce extremely advanced hardware, and the specific challenges those companies face offer the scientific community benefits that they won’t find with top of the line GPU’s.

The team continues their work as new developments in the hardware industry arrive, and one can find the full published results online.


Leave a Reply