ERCIM News No.26 - July 1996 - CLRC

The UK Collaborative Computational Projects

by Paul Durham

A problem which must be common to all ERCIM partners is that of gathering expertise together to address the problems of large scale software developments. One solution to this that has been applied in the UK is to establish Collaborative Computational Projects (CCPs). These have been sucessful both in completing the main software development targets, but also in technology transfer between organisations. At a time when much of the European Union is considering responses to Edith Cresson's Green Paper on Innovation (http://www. cordis.lu/cordis/grnpaper.html) CCPs provide a model which other countries may wish to follow.

The CCPs bring together all the major UK groups plus overseas collaborators in a given area of computational science to pool ideas and resources on large-scale software developments of mutual interest and general importance. This activity is on a scale larger than could be considered by a usual academic research group, and can receive additional funding from one of the UK governemnt research funding councils through a grant. Typically, CCPs:

Implement flagship code development projects
Maintain and distribute code libraries
Organise training in the use of codes
Hold meetings and workshops
Invite overseas researchers for lecture tours and collaborative visits
Issue regular newsletters.

The current CCPs are given in the following list:

CCP1 - The Electronic Structure of Molecules
CCP2 - Continuum States of Atoms and Molecules
CCP3 - Computational Studies of Surfaces
CCP4 - Protein Crystallography
CCP5 - Computer Simulation of Condensed Phases
CCP6 - Heavy Particle Dynamics
CCP7 - Analysis of Astronomical Spectra
CCP9 - Electronic Structure of Solids
CCP11 - Biosequence and Structure Analysis
CCP12 - High Performance Computational Engineering
CCP13 - Fibre Diffraction
CCP14 - Powder and Single Crystal Diffraction.

Each project has a Chairman and a Working Group which sets the scientific agenda, decides the work programme, and monitors progress. Contacts for these projects can be found at: http://www.dl.ac.uk/TCSC/projects/main.html

The entire CCP Programme involves about 240 academic groups in most UK universities, and around 150 collaborating groups in the rest of Europe, the USA, Japan etc. The list indicates the tremendous breadth of science and engineering covered by the CCP programme as a whole. This is not the place to describe the current interests of CCPs in detail, but much more information is available from me on request. Feature articles also appear from time to time in a publication from CLRC's Daresbury Laboratory, 'HPCProfile', copies of which can be obtained by contacting HPCProfile@dl.ac.uk. Daresbury Laboratory (DL) and Rutherford Appleton Laboratory (RAL) joined together to form CLRC, the computing departments of both DL and RAL are soon to merge into a single Computing department.

Typical resources available to a CCP include the running costs of organising meetings, newsletters, visits, etc. In addition, some CCPs have a post-doctoral research associate (PDRA) to carry out flagship code development; currently 5 PDRAs are in place. Finally, the CCP Programme is supported by a number of permanent CLRC staff. The standard funding mechanism is a 3-year Research Grant from an appropriate UK national research funding council.

The CCP programme as a whole is overseen by a Steering Panel comprising CCP chairmen and independent overseas members (most recently Giovanni Ciccotti and Roberto Car). Its main activities are to:

Provide a forum for discussion of general computational issues
Encourage inter-working of Projects, where appropriate
Review the work of permanent Daresbury staff
Encourage the development of new Projects.

How Successful are the CCPs?

To operate at a world-class level in computational science and engineering involves a major and long-term investment in software development and maintenance, dissemination and training. The key function of the CCPs is to provide a framework within which all the active groups in a given field can work together to make this happen. Achievements include:

Around 300 codes of varying sizes up to several hundred thousand lines of code are supported on many different types of hardware
High quality publication record; the last survey showed that the Projects produce an average of about 64 publications/year each, of which about 11% are in the Phys. Rev. Lett./Nature hot news category
Strong presence in 'Grand Challenge' Consortia within the UK's national High Performance Computing Initiative
Extensive participation in European programmes (HCM, TMR, ESF Networks)
Considerable interest from industrial research groups.

Moreover, it is worth noting the longevity of the CCPs. The first projects were initiated in the late 1970s and early 1980s, and have maintained their funding within rigorous peer-review processes which have themselves changed fairly frequently. A fascinating account of the birth of the first project, CCP1, can be found in Stephen Smith's and Brian Sutcliffe's article 'The Development of Computational Chemistry in the United Kingdom' to be published in 'Reviews of Computational Chemistry', volume 10 (1996), which also has many resonances for other areas of computational science and engineering. While a couple of projects have fallen by the wayside (CCP8 - Nuclear Structure; CCP10 - Plasma Physics), most have remained healthy, productive and valued, and new Projects have come through regularly.

The CCPs form an active, high quality, co-ordinated and cost-effective programme of computational science and engineering - a key element in the UK's support infrastructure for high performance computing applications.

Please contact:
Paul Durham - CLRC
Tel: + 49 1925 603263
Email: P.J.Durham@dl.ac.uk

return to the contents page