please note that we are currently updating this page
Finding ways to reduce the energy consumed by our computing devices whilst sustaining the performance levels, be they mobile phones, data centre servers or supercomputers is the most significant challenge the computer industry faces today. ICT has always been a rapidly evolving sector but significant changes in the IT market such as the increased uptake of storage and computing in the cloud and the maturing of other technologies such as data intensive analytics has led to an unprecedented increase in data volumes.
The management of data in all stages of its lifecycle - from creation through to destruction, and the need to analyse, integrate and share data often in real time - has driven a surge in demand from consumers and businesses alike and consequently increased the energy consumption associated with these services. The Energy Efficient Computing Research Programme has been established through a £19M capital grant from the Department of Business Innovation and Skills to establish a centre of best practice in the UK that will enable users of computer systems to achieve the same outcomes while minimising the consumption of energy.
Globally, there are three strands to energy efficient computing research; the development of energy efficient data centres, the development of energy efficient computers and the development of energy efficient software and algorithms. The EEC Programme seeks to draw together expertise in all three areas by developing our existing expertise and capabilities and through working collaboratively with other leading research centres and industrial organisations to pursue co-design efforts.
Our 'novel cooling demonstrator' comprises node boards immersed in mineral oil as coolant. Integrator: ClusterVision using GRC tank+oil and SuperMicro X9 boards comprising dual socket Intel IvyBridge
The Energy Efficiency Computing research group will work with manufacturers & industry to provide solutions that enable
Processor/chipset to run any given code with lower amounts of energy
Any given code to run with least amount of energy on any given platform
Every data centre to be more efficient in the running of user codes
Specific ideas include
showcasing hierarchical ability to measure and monitor energy consumed, from specific chip instructions right through to the data centre
methodology to predict energy consumed by specific codes, and by specific processor architectures
develop and promote techniques to reduce energy consumed by hardware, by software, and by efficient use of data centre resources
And as part of the bigger picture, EEC will contribute to Hartree's reputation by
Build up Hartree brand recognition as the go-to place for industry and SMEs who wish to tackle HPC,energy efficient parallel computing and Big Data analytics problems.
Develop a new range of joint Research and Development Projects, engaging in the active design of next generation energy efficient systems and software.
Establish Hartree Centre as a Centre of Excellence for applications development and optimisation for the next generation of low power processor technology such as ARM 64 v8 and the Knights (MIC) family of products.
Our dataflow cluster supports use of low power FPGA hardware for HPC. Comprises Maxeler Technologies DFE (data flow engines) and a Panasas parallel file system