Our computational research
Simulation of an amorphous hybrid metal-organic framework structure
Our overall research portfolio is underpinned by an impressive strength in computational and simulation research. This extends across a number of academic schools. Although the applications and methods in this work are very diverse, all research in this area relies heavily on the provision of large-scale computing power. Broadly this involves:
- a mixture of supercomputing or higher-capability facilities which harness the power of many processors working together via very fast data connections to solve complex problems
- high-capacity resources in which a single study involves huge numbers of separate simulations within a large parameter space
- analysis of multiple data streams which require large clusters of independent processors.
Materials and chemical simulations
Simulation methods are being used in a number of schools to understand the properties and behaviour of materials and molecules. A huge number of methods are used, as appropriate to the different types of study taking place.
Broadly speaking there are two ways to model the forces between atoms. The first is to use quantum mechanics directly (these are called ab initio methods), and the second is to use a numerical function with empirical parameters. Atomic scale processes within materials are studied using energy minimisation, lattice dynamics, Monte Carlo and molecular dynamics techniques.
Examples include studies of glasses and liquids, nanoparticles, structural phase transitions, anomalous properties such as negative thermal expansion, molecular crystals, catalysis and radiation damage in ceramics, metals and glasses.
Queen Mary is a partner in the Thomas Young Centre for Theory and Simulation of Materials.
Particle physics
Massive particle accelerators, particularly the Large Hadron Collider in CERN, generate vast quantities of data associated with the collisions of high-energy particles. Making sense of the data, and locating interesting events that will lead to major discoveries represents an unprecedented challenge. This challenge has spurred the development of a national and international computational grid – GridPP – in which Queen Mary scientists are playing a leading role.
Astrophysics
This image shows the vorticity distribution from a local 3D magnetohydrodynamic (MHD) simulation of a protoplanetary disc.
Computational methods are used in astrophysics research to simulate the behaviour of gas clouds and plasmas, with applications to planetary formation, collisions between planets, formation of accretion disks, and the behaviour of the solar plasma. Methods use both particulate and hydrodynamic approaches.
QM facilities
Queen Mary runs two facilities for both high-performance and high-capacity computational work:
- The high-performance facilities include two parallel computers with 1136 Intel cores and 868 AMD cores, both with inifiband interconnects
- The high-capacity facilities are associated with the GridPP cluster, and have 1440 cores.
Queen Mary scientists also use the HECToR national supercomputing facility.