HPCwire presents the inaugural edition of The Weekly Top Five. Our Friday wrapup (formerly The Week in Review) has been overhauled, and we hope you enjoy it. The new format features the five biggest HPC stories of the week, condensed for your reading pleasure.
This week, we cover the three-petabyte NEC storage installation at the French Brain & Spine Institute; the Platform HPC cloud survey; NVIDIA’s ARM-based CPU-GPU processor strategy; Glasgow’s thousand-core FPGA; and Whamcloud’s new talent.
NEC Provides Massive Storage System to French Brain & Spine Institute
The Institut du Cerveau et de la Moelle Épinière (ICM), the French Brain & Spine Institute, located in Paris, will now have a lot more space, storage space that is, 3 petabytes worth, provided by NEC and its partner Active Circle. Neuroscientists at the ICM together with over 600 researchers are engaged in developing treatments for major nervous system disorders. This important work requires a computer infrastructure that can manage immense amounts of data.
Bruno Lecointe, sales manager at NEC, highlighted an integral part of the deal: “One of the key objectives of the IT team at the ICM was to ensure that the technical infrastructure — and in particular data storage — should not be a hindrance to research. NEC, along with its partner Active Circle, enabled the IT team to provide a massive-capacity, three-petabyte-storage solution that meets this objective and is fully redundant, secure and cost-effective.”
Platform Survey Looks at HPC Cloud Experimentation
This week Platform Computing released the results of a survey that point to cloud computing’s growing acceptance from the high performance computing (HPC) community. Out of the 100 participants surveyed during SC10 in November 2010, sixty-two percent of users had experimented with private or public clouds, with nearly half those respondents reporting a positive experience.
Randy Clark, CMO, Platform Computing, opines: “Until recently, HPC users have been skeptical of implementing HPC cloud environments due to potential performance issues stemming from virtualization and the need for proper load balancing and fast compute times. The data we collected from delegates at this year’s Supercomputing Conference show that the tide is turning for private cloud within the HPC community. Our customers are also reflecting this trend with more beginning to explore capabilities that help them better manage workloads while keeping an eye on their costs.”
NVIDIA’s Unveils ‘Project Denver’ at CES
At the Consumer Electronics Show (CES) in Las Vegas, NVIDIA announced it is licensing the ARM architecture in order to build a new class of processor, one that marries CPU to GPU. An NVIDIA CPU running the ARM instruction set will be fully integrated on the same chip as the NVIDIA GPU. “Project Denver,” as the initiative is known, is the result of a strategic partnership that gives NVIDIA the right to develop CPU cores based on ARM’s future processor architecture. The new ARM-based processors will power future generations of personal computers, servers, workstations and supercomputers.
In the company’s press release, Jen-Hsun Huang, president and chief executive officer of NVIDIA, states that Project Denver “marks the beginning of the Internet Everywhere era, where every device provides instant access to the Internet, using advanced CPU cores and rich operating systems.” He remarked that “ARM’s pervasiveness and open business model make it the perfect architecture for this new era.”
HPCwire has further coverage of NVIDIA’s plans here.
Glasgow Scientists Develop Thousand-Core FPGA
Glasgow-based researchers are looking into the viability of using the Field Programmable Gate Array (FPGA) processor as a type of manycore device. Dr. Wim Vanderbauwhede of the University of Glasgow’s Department of Computing Science in partnership with colleagues at the University of Massachusetts Lowell have modified an FPGA processor so that it effectively contains more than a thousand cores.
From the release:
FPGAs can be configured into specific circuits by the user, rather than their function being set at a factory, which enabled Dr Vanderbauwhede to divide up the transistors within the chip into small groups and ask each to perform a different task.
By creating more than 1,000 mini-circuits within the FPGA chip, the researchers effectively turned the chip into a 1,000-core processor — each core working on its own instructions.
To test the chip’s compatibility with real-world applications, the team fed it an algorithm used in the MPEG movie format — a YouTube video standard. The divied-up FPGA was able to process the algorithm at a speed of five gigabytes per second, around 20 times faster than what a top-end desktop might muster.
Key Lustre Engineers Make Leap to Whamcloud
Whamcloud just significantly increased its talent pool with the hiring of three top Lustre engineers from Oracle. The new employees — Bryon Neitzel, Andreas Dilger and Peter Jones — will start this month, continuing to champion Lustre in their management, community and support roles.
Says Whamcloud CEO and President, Brent Gorda: “We’re bullish about where Lustre is headed. Recent attention from the community shows there is more interest and support than ever in moving the technology forward. We want to be the place people think of – worldwide – when they think of Lustre.”