Running code on graphics cards can boost application performance tenfold
Scientists have developed a method of speeding up applications 10 times by transforming the underlying code to run on a graphics card.
Researchers from the University of Bedfordshire Centre for Computer Graphics and Visualisation, alongside another team based at the University of Groningen in the Netherlands, have created a toolkit for “GPUtilisation” in SME applications (GPSME). The GPSME Toolkit is designed to be used by developers to convert code that would normally run on a computer’s CPU into a format that can be understood by a graphics processing unit. The team claimed this transformation takes minutes to complete.
It does compute
Such a conversion would allow computers to “perform at a higher level”, the researchers claimed, as a GPU system can manage thousands of tasks at a time. Four small companies in Europe, including London-based software firm AnSmart, have so far tested the trial web service, which is due to be launched in September.
“The GPSME Toolkit is unique because it allows software designers – or those with software codes of computer applications – and SMEs, without an in-depth understanding of GPUs, to semi-automatically transfer codes from CPU to GPU,” said Feng Dong, professor of visual computing at the university. The toolkit website has easy-to-follow tutorials to help users transfer a wide-range of computer programs so that they now run off the graphics card rather than the CPU, which as many people are now aware enables PCs to run faster.”
He also said the service would benefit SMEs facing ever-increasing competition, as it would allow small firms to improve their products in terms of speed and quality without major overheads.
“The recent explosion of GPU power has not been fully utilised by many SMEs, possibly because GPU programming requires specialist skills different from those of conventional programming. GPSME will provide the SME participants with a simple route to accessing GPU power,” Dong said.
The toolkit will also support the execution of advanced techniques within acceptable runtimes and “hence allow the SMEs to use more complex computing models in their new products,” Dong claimed.
He added that this would bring SMEs major commercial benefits and “significantly improve their market positions”.