CCP3 - Computational Materials
The Computational Materials Group at Daresbury supports CCP3, the Collaborative Computational Project for simulation of physical and electronic properties of surfaces and interfaces, see CCP Web site.
The group is involved in the development of the CRYSTAL, CASTEP and DL_EXCURV codes and DLV as a GUI for visualisation and supports work on experimental facilities such as SRS, ISIS, MEIS and Diamond as well as the UKCP Materials HPC Consortium.
LighTneT is a major new collaboration on theoretical support for light source facilities (synchrotrons) which has been established and funded by the EU at the level of 1 million Euro's. This represents an International outreach of CCP3's activities Collaboration between DL/ ESRF/ Trieste/ INFN and to include Diamond in the future.
DLV 3.0 provides an interface to the HPC codes and is based on AVS Express 5.0 and written in C++. It has been tested on the e-Science Centre SCARF cluster and is being implemented on NW-GRid using the GROWL client toolkit.
Science Testbed Runs
Typical calculations, which are ideal for Grid clusters such as NW-GRID include:
- Systems involving 50-250 atoms
- Scale efficiently to around 8-32 processors
- Geometry Optimisation
- o Interested in parallel methods
- Calculation of Phonon modes
- o weakly coupled parallelisation
- Transition state searches
- o weakly coupled parallelisation
FUNFLUOS Project Testbed
Project Investigators
C.L. Bailey, S. Mukhopadhyay, A. Wander and N.M. Harrison
Scientific/ Technical Objectives
To use ab initio methods to characterise the surface structure and properties of aluminium fluorides.
Role of NW-GRID
This project requires a large amount of processor time. We typically run geometry optimisation calculations in parallel on up to 32 processors. A single calculation typically takes around 60 hours. We have been able to use NW-Grid to run a significant proportion of these calculations. For example, we have calculated the surface structure of alpha-AlF_3 as a function of HF and H_2O partial pressure. The resultant phase plot for the (001) surface of alpha-AlF_3 is shown below. This study involved over 150 geometry optimisations using hybrid exchange density functional theory as implemented in the CRYSTAL code.
Applications Software
CRYSTAL
Grid Software
GLOBUS
Progress to date
The structure of the (001) and (012) alpha AlF_3 surfaces have been calculated as a function of HF and H_2O partial pressure. We are currently characterising the surface of beta AlF_3. This surface is known to contain Lewis acidic catalytic sites and has potential as an industrial catalyst. We have calculated binding energies and vibrational spectra of NH_3 and CO molecules adsorbed onto these sites in an attempt to further understand the catalytic properties of the surface.
Publications
S. Mukhopadhyay, C.L.Bailey, A. Wanderr, B.G. Searle, S.L.M. Schroeder, R. Lindsay and N.M. Harrison, accepted in Surface Science (2007)
Future Policy Requirements
During this work the following requirements were identified:
- CPU Counts - May require relatively low processor counts (10's
- rather then 100's) for long run times (48 hour plus). Complimentary to HPCx resources;
- Job Submission - Ability to submit to single queue - resource
- discovery is important;
- Run weakly coupled jobs - many compute intensive jobs with low
- communications needs;
- Flexibility - queuing policy needs to be flexible during testing
- and design;
- Interactive Queue - must have the ability to quickly run short
- debugging and development jobs;