Service Communautaire d'Information sur la Recherche et le Développement - CORDIS


GPSME Résumé de rapport

Project ID: 286545
Financé au titre de: FP7-SME
Pays: United Kingdom

Final Report Summary - GPSME (A General Toolkit for “GPUtilisation” in SME Applications)

Executive Summary:

SMEs, particularly those whose business is focused on developing innovative products, are subject to many pressures in maintaining and growing their market share and ensuring that their products remain competitive in an age of rapid technological change. In many high-tech fields, users are experiencing a huge growth in data, with increases in quantity, in resolution, in variety, etc., while the work often presents significant time constraints on the associated data processing. This leads to a continual upward pressure on computational resources. The initiative of this project came from the demands of the 4 SME participants. While providing services in different areas, they all face a common problem: the quality of their products has been inhibited by a lack of computing power.

Parallel computing offers fast computing by splitting tasks into small components and distributing them among multiple processors/threads. The remarkably increased power of Graphics Processing Unit (GPU) in recent years offers a very attractive alternative, which can handle many demanding tasks by only harnessing local computing resource in low-cost computer platforms.

GPSME provides the SME participants with a simple route to accessing GPU power. By involving close cooperation between the SMEs and RTD performers, GPSME develops a toolkit to automate the conversion of existing sequential CPU code to an optimal GPU implementation. The toolkit allows the SME users to improve the speed performance of their products quickly and economically by automatically identifying the parallelizable sections of their programs and converting them into GPU implementations. With such a toolkit, the SMEs are able to convert their existing CPU code without committing significant effort and time. It also supports the execution of advanced techniques within acceptable runtimes and hence allows the SMEs to use more complex computing models in their new products. This brings them major commercial benefits and significantly improves their market positions.

Technically, GPSME features techniques to adapt automatic parallelization to the latest GPU compute architecture to deliver optimal performance. This is expected to greatly improve on traditional CPU-based automatic parallelization. The literature review has suggested that the techniques in this area are still very much in their infancy and that there is no existing toolkit that can benefit the SMEs immediately.

The application areas of the SME participants are widely differing, so GPSME produces a breakthrough that opens the door for great performance gains across many areas of application. This development is particularly suitable for SMEs that focus applications on moderate platforms. In the long term, the outcomes of GPSME benefits many companies and improve industrial competitiveness across the whole European Union.

To achieve this aim, the specific objectives of GPSME and measurable outcomes are given below:

I.Techniques allowing adequate adaptation of automatic parallelization to GPUs by taking full consideration of GPU compute architectures with a range of compute capabilities.
II.A toolkit that performs GPUification suitable to the SME application areas.
III. Scientific validation of the GPSME toolkit in the application areas.
IV.Enhanced speed and computing precision in each of the SME applications.

Project Context and Objectives:

Please see the attached document - Final_report_GPSME_Publishable_Summary.pdf.

Project Results:

Please see the attached document - Final_report_GPSME_Publishable_Summary.pdf

Potential Impact:

Please see the attached document - Final_report_GPSME_Publishable_Summary.pdf

List of Websites:

Informations connexes


Feng Dong, (Professor of Visual Computing)
Tél.: +44 1582 743940