RealityGrid

RealityGrid

Peter Coveney1 and John Brooke2

1. Centre for Computational Science, Department of Chemistry, Queen Mary,

University of London

2. Manchester Research Centre for Computational Science, University of

Manchester

2

The Problem

• Simulation time: days. Analysis time: months

• Ability to generate data outstrips our ability to understand it by several orders of magnitude

3

RealityGrid Aim

Using grid technology to closely couple high performance computing, high throughput experiment and visualization,

RealityGrid will move the bottleneck out of the hardware and back into the human mind.

4

E-Science Framework

Code Code

meta meta

meta meta

Verification,Optimisation &

Scheduling

Visual ProgrammingEnvironment

Code Code

Distributed Component Repository

Globus ExecutionEnvironment

Meta-data: Software interface and component performance

Performance feedback

through GrADS & APART

into meta-data

Layered ArchitectureApplications / Problem Solving Environments

Grid ServicesHBMGASS

Grid FabricLSF

MPI

NQE

Application ToolkitsGlobusView

Solaris

GSI-FTPMDS

Grid Resources

Linux

PBS

GSI GRAM

DUROCMPICH-G globusrun

Manchester Imperial College

EPCCOxford

QMLoughborough

Manchester

QM-LUSI/XMT

UNICOS IRIXTru64

SRB

LUSI Portal Component RepositoryVisualization & SteeringComputational PSE

Component FrameworkVIPAR

6

Integrated experiments

• Scientific discovery can be enhanced by closely coupling computation and experiment

• RealityGrid includes two such experimental components: – X-ray microtomography : produces 3D X-ray attenuation

maps of specimens at a microscopic level– London University Search Instrument (LUSI) a materials

combinatorial science laboratory for study of ceramic materials

7

X-ray Microtomography• Simulation, visualization

and data gathering coupled via RealityGrid

• Expensive synchrotron beam time resources optimally used to obtain sufficient resolution for simulation

• Local testbed providing grid enablement model for European synchrotron facility

8

LUSI firsts

• Almost all drug design is now done in a combinatorial manner

• LUSI is the first instrument to apply this technique to materials science problems.

• LUSI provides high volume experimental data directly to the high performance analysis tools enabled by the grid.

• Grid enabling LUSI will allow researchers from other institutions to add to and mine from its unique database.