EC touts ‘data centre in a rack’ tech to tackle HPC efficiency problems

EC touts ‘data centre in a rack’ tech to tackle HPC efficiency problems

A new programme will develop open source designs of computing building blocks

Article comments

The European Commission has launched a programme, dubbed CoolEmAll, to tackle the energy efficiency implications that are likely to arise from increased investment in High Performance Computing (HPC) infrastructure. 

Exa-scale computing is used in a variety of industries to carry out rapid simulations and the EC has announced plans this month to double its investment in HPC from €630 (£532) to €1.2bn (£1.01bn) before 2020. 

However, HPC requires a datacentre infrastructure that consumes a large amount of energy and resources. For example, one supercomputer can consume more power than a medium sized data centre and can cost more than $100,000 (£63,700) a year to run. 

CoolEmAll, a consortium of companies and organisations, plan to tackle this problem by developing two key tools to help industries better monitor and manage their supercomputer energy consumption. 

The main outcome of the project will be a set of open source designs based on a high density server called the RECS Compute Box, which is being developed by a German start-up called Christmann Informationstechnik. 

The RECS Compute Box has been hailed as a “data centre in a rack”, where it consists of several high-density server units, each containing up to 18 CPU-Modules, which means that it can contain up to 600 individual server boards. This is approximately 10 times what you would typically expect to find in a traditional server rack. 

Each Compute Box is also contains 18 thermal and 18 current sensors integrated into the baseboard infrastructure, which enables the network status, fan speed and power usage of each of the 18 boards to be monitored. 

The EC is hoping that the designs of these high density, high efficiency servers will be used commercially for HPC projects in Europe in the future. 

Alongside the hardware, CoolEmAll is also looking at developing simulation, visualisation and decision support software to help provide a real-time modelling tool for companies using HPC. It is hoped that this will allow datacentre planners to model the energy efficiency implications of where servers are physically placed, cooling measures, as well as the role played by applications. 

The CoolEmAll project is made up of seven HPC research organisations, including The 451 Group, Poznan Supercomputing and Networking Centre and the High Performance Computing Centre University of Stuttgart. 

Andrew Donoghue, analyst at The 451 Group, confirmed that the first developments from the CoolEmAll project are likely to be presented to the European Commission before the end of 2012. 

Share:

Comments

Advertisement
Send to a friend

Email this article to a friend or colleague:


PLEASE NOTE: Your name is used only to let the recipient know who sent the story, and in case of transmission error. Both your name and the recipient's name and address will not be used for any other purpose.


We use cookies to provide you with a better experience. If you continue to use this site, we'll assume you're happy with this. Alternatively, click here to find out how to manage these cookies

hide cookie message

ComputerworldUK Knowledge Vault

ComputerworldUK
Share
x
Open
* *