Grid computing at RHUL

Access to massive computing resources is becoming key to doing particle physics research. The same trend is emerging across science in general. We will need a huge amount of computing power and data storage capacity to analyse data from the ATLAS and the other LHC experiments. To this end, we operate three high performance computing clusters: Faraday, Newton and Rutherford.

Our latest addition in April 2012 takes it up to a total 1750 logical processor cores and 700 Tera Bytes of RAID6 disk storage.

Our second acquisition 'Rutherford' in March 2011 has 480 processor cores and 420 Tera Bytes of RAID6 disk storage. It was supplied by Dell and integrated by Alces.

Our older cluster 'Newton' has 960 kSI2k of processing power from 400 processor cores and 300 Tera Bytes of RAID6 disk storage. The Newton cluster was supplied by Clustervision.

Both Newton & Rutherford the gLite middleware to provide Grid compute and storage services to a wide range of virtual organisations, with particular emphasis on the ATLAS experiment at the LHC.

These clusters form part of the London Grid, a consortium of London colleges which provides Grid computing services to the UK particle physics Grid, GridPP. The primary aim of the Grid service is to deliver the computing facilities needed to analyse LHC data. However, during commissioning, spare processing time has been put to good use on projects like simulations to test potential Avian flu drugs.

Contact: Pedro Teixeira-Dias, SimonGeorge.