HPC news for supercomputing professionals

Entries filed under “Green HPC”

Design and management techniques that contribute to the responsible, effective use of energy in the operation of high performance computing centers and equipment.

Keeping up with the Googles

You might have a big datacenter. You might even have a top 20 supercomputer. But are you negotiating 20-year energy leases with windfarms in Iowa? I didn’t think so, poser.

In a continuing effort to find new ways of greening our operations, we will begin purchasing the clean energy from 114 megawatts of wind generation at the NextEra Energy Resources Story County II facility in Story and Hardin counties in Iowa on July 30 at a predetermined rate for 20 years

To be fair, Google isn’t buying the energy for its direct consumption. They intend to sell it on the spot market for energy, and are hoping that the long term contract will help stimulate larger investments in renewable …

1 Comment

The Economist spots wave of innovation in cooling systems

The Economist is reporting on a trend toward more innovation in the air conditioning business driven by energy costs and the desire to be more environmentally responsible with the energy we do use. Some ideas the article calls out? Blowing air over ice, thermal coolers that use waste hot water to cool spaces, and (in dryer parts of the planet) evaporative coolers.

And then there’s this

However, researchers at the National Renewable Energy Laboratory (NREL) in Colorado have designed an evaporative system that sprays ambient-temperature water into warm air to cool it, but in a way that also lowers the humidity. NREL uses syrupy liquids which contain salty desiccants to soak up the humidity. Hot water is used to heat the syrups and dry them

Also posted in Datacenter operations | 1 Comment

IBM, accelerators dominate Green500

The latest Green500 list was released at the end of June, and once again IBM dominates the list with a range of platforms.

IBM is in 17 of the top 20 slots, and 65 of the top 100, with Cell, iDataPlex, and Blue Gene. The number 1 system on the list, an IBM at Forschungszentrum Juelich (FZJ), is ranked with 773.38 MFLOPS/W, and shows up at #5 on the Top500 list. For comparison, the number one on the Top500 (ORNL’s Cray Jaguar system) is 56 on the Green500 with only 253 MFLOPS/W. Which just goes to show that it’s all about what you measure.

As the Green500 web site points out, 8 of the top 10 systems are accelerator-based (6 use …

7 Comments

HP publishes first TPC-Energy results

Today the Transaction Processing Council announced that HP submitted the first published results for the new TPC-Energy benchmark, a set of measures that augment the existing TPC-C, TPC-E and TPC-H benchmarks (we talked about TPC-Energy here).

The Transaction Processing Performance Council (TPC) today announced the first results for its TPC-Energy specification. Hewlett-Packard Company has published TPC-Energy results on all three TPC benchmarks: TPC-E, which simulates the online transaction processing (OLTP) workload of a brokerage firm; TPC-C, which simulates the OLTP order-entry workload; and TPC-H, which simulates a decision support workload. Each of these benchmark publications also includes the optional Watts per performance metrics.

The TPC isn’t the only bear in the energy benchmark woods, though; the Standard Performance Evaluation Corp. (SPEC) has been …

Also posted in Datacenter operations | 2 Comments

IBM’s Aquasar-cooled system goes live in Zurich

In June of last year we wrote about a new water-cooled system that the Swiss Federal Institute of Technology Zurich [ETH] and IBM were working on. The new system is called Aquasar, and will directly repurposes excess heat from the computer for university buildings.

According to news at Data Center Knowledge, the 6TF system is now live

The deployment comes a year after ETH and IBM scientists announced plans to collaborate on chip-level water-cooling and energy reuse. The supercomputer, which is now fully operational, consists of special water-cooled IBM BladeCenter Servers designed by IBM scientists in Zurich and Boblingen, Germany. Each of the blades will be equipped with a high-performance liquid cooler on each processor, as well as input and output

Also posted in Datacenter operations, New Installations | Leave a comment

ARM funded by EU to build low power datacenter

Last week ARM Holdings, the Cambridge, UK based company that designs the well-known ARM processor, announced that it is leading a project to prove out the idea that its low-power processors can power datacenter-sized computations

ARM today announced it is leading a Green Cloud Services project called EuroCloud, along with Nokia, IMEC, EPFL, and the University of Cyprus. The research project, which is funded by the European Commission FP7 Programme for an initial three years, aims to validate the value proposition of constructing data center platforms using a combination of low power processing and 3D chip packaging solutions to deliver a significant improvement in cost- and energy-efficiency compared to the approach utilized in incumbent server platforms.

…This research project

Also posted in Business of HPC, Cloud HPC | Leave a comment

Concurrent Thinking spins out, raises 1.5M USD

Concurrent Thinking, a UK-based HPC company, has announced that it has spun out of its parent company and secured a round of initial funding to help with the change

Concurrent Thinking Ltd today announced that it had spun out its cluster appliance business into a new company. The new company will exploit technology that was conceived and sold by Streamline Computing Ltd, the UK’s leading integrator of High Performance Computing (HPC) systems. Over the last 18 months, this technology has been deployed at more than 40 of Streamline’s customers in the UK.

The spin-out follows the successful completion of a syndicated funding round of over £1M, which included original shareholders as well as a new investor, Carbon Trust Investment Partners. The funds will be used

Also posted in Business of HPC | 1 Comment

US Senate considers bill to improve energy efficiency, data centers not mentioned

S.3251, “Improving Energy Efficiency and Renewable Energy Use By Federal Agencies Act of 2010″, is scheduled to get its first major action of this Congress on June 15 in hearings scheduled by the Senate Subcommittee on Energy. The bill, introduced in April by Thomas Carper (D-DE), includes provisions that mandate the use of “computer hardware, energy efficiency software, and power management tools” to reduce power consumption in the government’s fleet of personal computers.

That’s right: SEC. 6. ADOPTION OF PERSONAL COMPUTER POWER SAVINGS TECHNIQUES BY FEDERAL AGENCIES only mentions personal computers, not datacenters. We do know that the Congress is taking an interest in the data center plans of individual agencies, and you still have …

Also posted in Datacenter operations, National and Legislative Action | 2 Comments

EPA finally unwraps labeling program for data centers

This week the EPA formally initiated its long-discussed Energy Star labeling program for data centers and the buildings that house data centers.

To earn the label, data centers must be in the top 25 percent of their peers in energy efficiency according to EPA’s energy performance scale. By improving efficiency, centers can save energy and money and help fight climate change.

EPA uses a commonly accepted measure for energy efficiency, the Power Usage Effectiveness metric, to determine whether a data center qualifies for the Energy Star label. Before being awarded the Energy Star, a licensed professional must independently verify the energy performance of these buildings and sign and seal the application document that is sent to

Also posted in Datacenter operations | 2 Comments

HP expands datacenter location checklist to include ready source of manure

I was going to use “poop” in the title in honor of my kids, but then I thought that was unprofessional. Then I thought, “hey, this is a blog.” As you can see I went with a hybrid approach…thoughts?

Anywho, c|net’s Erica Ogg reported earlier this week that researchers from HP Labs are presenting a paper at the ASME International Conference on Energy Sustainability conference on the use of waste products from cows to provide power for datacenters

As odd as it may sound, that’s where the cows come in. The average dairy cow produces 55 kilograms of manure per day, or 20 metric tons per year. An individual cow’s manure can generate 3 kilowatt-hours of

Also posted in Datacenter operations | 1 Comment

Podcast Series

Green HPC Podcast

insideHPC.com is a production of indigoBit, LLC. © 2006-2010 Sitemap