December 9, 2010

The Week in Review

by Tiffany Trader

Here is a collection of highlights from this week’s news stream as reported by HPCwire.

ORNL Supercomputing Research Opens Doors for Drug Discovery

Ethernet Vendors Demo’ed 10, 40, 100 GbE Interoperability at SC10

CD-adapco, R Systems Help Wilson Improve Football Design

Purdue Computing Resources Rank Near Top Nationally

Report Finds K-12 Computer Science Education Declining

Lockheed Martin Selects Appro CPU-GPU System for DoD Work

Whamcloud Announces Lustre Contract with ORNL

Exludus Launches Application Profiling and Reporting Solution

AMAX Adds GPGPU Blade Cluster Offering

OCF Announces Cluster On Demand Service

U of Colorado’s SC10 Cluster Competition Team Receives Fan Favorite Award

Altera Unveils Process Technology Strategy for 28-nm Product Portfolio

Xilinx Extends Comprehensive DSP Portfolio with New Development Kits

Panasas Strengthens Executive Leadership Team

DDN Expands Global Footprint and Leadership Across Listing of World’s Fastest Supercomputers

Berkeley Lab Scientists Receive Time on Nation’s Fastest Computer for Energy Research

UK Reaches Out to US for Supercomputing Collaboration

A UK-based development project, The North England Inward Investment Agency, is seeking technological insight across the pond. Specifically the group is brokering relationships between US-based interests and Daresbury Labs in Northern England for collaboration on leading-edge supercomputing software projects. The overarching goal of the project is to create high-level software capable of extracting the full potential from today’s supercomputers.

John Bancroft is the head of business development for the Science & Technologies Facility Council (STFC) located in the North of England. Bancroft has been visiting sites in the US, including SC10 in New Orleans, to identify and develop strategic partnerships for the Hartee Research Centre, an extension of Daresbury Labs that will be working on “grand challenge” projects.

Stated Bancroft:

“Computers are getting more and more powerful but the software hasn’t kept up. Until the software is developed, the power of the new computers won’t be realized. Collaboration between hardware and software developers focused on high end computing is key to creating an environment where research is not based solely on academic curiosity but is firmly tied to application and economic benefit.”

Kristen Hirst of the North of England Agency discussed society’s pressing need for supercomputing software:

“Supercomputers are no longer things that we talk about affecting us at some undefined time in the future. They are here now and will most definitely be a critical component in the development of new drugs, new sources of energy and environmental modeling. What we need now is software designed to tap into that power.”

According to the release, American companies of all sizes are teaming with STFC’s Daresbury campus to pool their knowledge and skills. Putting together the US and European computational science capabilities creates a win-win synergy for both groups. To help open this opportunity up to smaller companies, Daresbury is offering free office space for those meeting certain requirements.

TACC’s FutureGrid Cluster Under New Management

The Texas Advanced Computing Center (TACC) has selected the Bright Cluster Manager to manage its HPC cluster in the FutureGrid project. FutureGrid is a distributed testbed that allows researchers to collaboratively develop and test new approaches to parallel, grid and cloud computing. Currently there are six clusters that comprise FutureGrid and TACC is hosting one of those, as well as contributing to the development of the user portal, and providing tools to assist with experients.

The director of Advanced Computing Systems at TACC, Tommy Minyard, commented on the selection process:

“TACC always evaluates several vendors, and for this particular project Bright was the best choice. We were impressed with the Bright Cluster Manager demo at the SC09 conference and decided to make a detailed evaluation here at TACC. Our findings confirmed to us that Bright was what we needed: a powerful cluster management solution that is easy to use, flexible and very capable. Bright met our demanding requirements straight out of the box.”

The FutureGrid project is lead by Indiana University and will be integrated into the NSF’s TeraGrid cyberinfrastructure. The testbed is enabling scientists to conduct experimental computational science research that requires specific configurations and repeatable experiments. The ultimate goal is for researchers to be able conduct compute-heavy experiments without having to worry about the underlying hardware design and capabilities.

Still want more information on FutureGrid? Indiana University has a Webpage dedicated to project highlights, including status updates.

Share This