Parallel computing optimises desktops
Parallel computing means to divide computer tasks among two or more processing chips. That gives a tremendous speed improvement compared to single-chip machines. The EU-funded project 'A general toolkit for GPUtilisation in SME applications' (GPSME) brought super-computing concepts to the office. Modern computers generally have an under-utilised second processor in the form of the graphics processing unit (GPU). The eight-member project aimed to network the GPU with the main processor. GPSME ran for two years to the end of September 2013. GPSME developed and tested tools for converting existing application code to utilise both processors. The resulting project software automatically recognises sections of application code suitable for conversion, and efficiently carries out the conversion. GPSME's work improves conventional automatic parallel conversion methods. The result is high computer performance at minimal cost. Presently, no other system is available to assist SME competitiveness in this way. Thanks to GPSME, European SMEs will have the resources they need to develop new products. That will mean opportunity and improved competiveness for European business.
Keywords
Computer performance, SMEs, parallel processor, graphics processing unit, application code
