Advanced Research Computing

Site Under Construction !!!

Computing Services for Researchers

The Advanced Research Computing (ARC) team provides a variety of services and resources aimed specifically at researchers including:

The ARC team have helped researchers across a wide range of disciplines to tackle their computing problems including physical sciences, engineering and health and life sciences.

If you are unsure of how to proceed or just want some general advice, contact the ARC team at: arc-support<@>liverpool<.>ac<.>uk.

High Performance Computing (HPC)

The main HPC resource, chadwick, is a Bull supercomputer dedicated to research use. The system is ideal for problems involving the solution of large systems of differential/integral equations found in (for example) Finite Element Modelling and many other application areas.

The hardware consists of 118 nodes , each of which are powerful computers in themselves. Nodes contain 16 cores and 64 GB of memory. A large memory node comprising 128 cores and 2 TB of memory is also available. In addition there is 48 TB of networked storage plus another 90 TB of parallel (fast access) storage available.

A wide variety of software packages are already installed on chadwick including VASP, CP2K, Abaqus, Fluent, MATLAB and R. Requests for additional software to be installed are viewed sympathetically, particularly if the software is free.

All Liverpool users also have free access to HPC systems provided by the N8 collaboration of Northern England universities. This includes a similar system to chadwick with 332 x 16 core nodes. Details can be found at: n8hpc.org.uk

Condor High Throughput Computing (HTC)

The ARC Condor Pool is not a dedicated system but comprises around 400 teaching/learning centre PCs located across the campus. Each PC has (effectively) a four core 3.3 GHz processor and 8 GB of RAM.

Users of the Condor Pool generally write their own codes rather than using third party applications and tools have been developed to aid in the running of codes written in MATLAB and R. (both of which ARC have significant experience of).

The Condor Pool It is is especially suited to running very large numbers of independent programs (jobs) where each runs only for a short time (20-30 minutes is ideal) however, it is also possible to run programs for arbitrarily long periods using a technique called check-pointing.

Condor can process large datasets quickly and help speed up statistically-based problems such as simulations. A good example of this is problems involving Monte Carlo methods. Condor is also very effective in developing and verifying statistical models coded in R.

Researchers in the health and life sciences (amongst others) have found that Condor has reduced time-to-solution from months to days and many Condor applications start life just as simple R (or MATLAB) codes run on a PC. You can find more about Condor on the ARC website at: condor.liv.ac.uk.

How to register

ARC services such as Condor and chadwick are provided completely free of charge to staff (including associates) and postgraduate students but we do ask that all users register for each service using ServiceNow - click:

Make a request > Accounts > Application to access high performance/throughput computing facilities