Dartmouth researcher building a cost-effective supercomputer using smartphones
Cheap and supercomputer are not normally used in the same sentence. However, thanks to a University of Massachusetts, Dartmouth physics professor, supercomputing is fast becoming affordable.
The professor, Dr. Gaurav Khanna, also associate director of the Center for Scientific Computing and Visualization Research (CSCVR), back in 2007, figured out how to turn PlayStation 3 hardware into PS3 Gravity Grid, a powerful, working supercomputer that cost considerably less than current supercomputers of equivalent capacity. In 2014, Dr. Khanna upped the ante, building a next-generation supercomputer called PS3 Reefer.
Dr. Khanna and PS3 Reefer
Image courtesy of UMass Dartmouth
“UMass Dartmouth has built an extremely low-cost supercomputer using 176 Sony PS3 gaming consoles installed in a refrigerated shipping container or ‘reefer’ of large cooling capability located conveniently on campus,” from a June 2014 UMass Dartmouth press release. “The cost effectiveness of this novel approach to supercomputing is nearly an order-of-magnitude greater than traditional supercomputers in data centers. This system’s performance is comparable to nearly 3000 processor cores of a typical laptop or desktop.”
If there is any thought these projects are gimmicks, that’s so not the case. In 2009, Dr. Khanna co-published two papers made possible by the PS3 Gravity Grid: one paper in the journal Parallel and Distributed Computing and Systems discussed how cell processors in PS3 consoles improved calculation speeds by a factor of 10, and a paper in the journal Classical and Quantum Gravity explained gravitational wave properties produced when two black holes merged.
As for PS3 Reefer, it is currently performing complex calculations dealing with black hole astrophysics, and vulnerabilities in cybersecurity.
Low cost, efficient, yet powerful
One may ask why Dr. Khanna, an astrophysicist, is messing around with PS3 gaming consoles. “It is about making supercomputing more accessible to scientists, mainly through offering cost-effective alternatives,” said Dr. Khanna.
For his latest supercomputing project, the professor added a new constraint. “Supercomputers need to be more power efficient and perform well using much less electricity,” explained Dr. Khanna in an August 2015 UMass Dartmouth press release.
To that end, and with some urging from Irwin Jacobs, cofounder of Qualcomm and developer of the successful Snapdragon cell phone processor, Dr. Khanna and his associates decided to look at smartphone processors.
It appears that initial testing is positive. From the press release: “Dr. Khanna and CSCVR researchers have performed initial tests on that idea. They were able to discover that if a supercomputer was built using mobile phone chips it would use 30 times less electricity for the same performance compared to traditional supercomputer servers.”
Next Dr. Khanna and fellow Dartmouth researchers employed Inforce Computing’s SBC to evaluate Snapdragon’s ability to run full-scale astrophysics and computational mathematics research codes. “Dr. Khanna’s team cautions that they have only tested a single chip,” explained the press release. “However, this (power efficiency) is a very positive sign in the early testing stages with a potential for huge savings on operating costs.”
With his track record, I expect to report in the not too distant future that researchers at UMass Dartmouth are using a smartphone supercomputer to figure out some strange astrophysical phenomena.
An interesting and possibly insightful conclusion by Dr. Khanna:
“It is well known that if we attempt to build the next generation supercomputer using today’s technology, we will need multiple nuclear power stations to simply turn it on. Power efficiency is the key in the future of supercomputing and that is why I am convinced that the next generation machines will be built using mobile phone parts.”