Monthly Archives: January 2007

IBM and Softek announced on Monday that IBM has acquired the 140-person data storage company for an undisclosed sum.

IBM officials say they want to integrate Softek’s data mobility technology with IBM’s expertise in storage and data services. The companies have been partners since 1996.

Storage is another murky backwater in HPC. We’ve spent all our time over the past several decades getting more and more bits into smaller spaces. But we’ve spent almost no time …

via Sun’s HPC Watercooler
Interesting pointer to an article in Gridtoday by Wolfgang Gentzsch about the state of “non-trivial” HPC grids:

A lot of the grid middleware currently promoted is really intended for research and demonstrations but needs significant effort to be made suitable for large-scale production usage.

From the “Hey, I have some of that same gear” department…

SGI issued a release today talking about the additional investment that EFILM has made in SGI gear to support the digital workflow it uses to create movie trailers for Hollywood.

To create the SGI SAN to provide high-speed storage and networking to run their new trailer business, EFILM purchased an SGI® Altix® 350 server with 8GB RAM and 2 Dual-Core Intel®

…

OSC issued a press release yesterday announcing that they had upgraded one of their educational clusters. The system is used for both visualization and computation.

What’s interesting about the press release is that they go ahead and factor the GPUs in when talking about performance:

The updated cluster boasts 55 rack-mounted compute nodes. Each node contains a dual core AMD Athlon 64 processor integrated with nVIDIA GeForce 6150 graphics processing units (GPU). An application taking full

…

ClusterMonkey has an interesting case study that examines what kind of cluster your could build today for $2,500 versus what you could do (and what they did) 18 months ago.

It’s a little long, but an interesting study in the design tradeoffs that lead to a balanced system at any scale.

But more importantly, I’ve learned that in 18 months technology has really changed. For $2,500 we have gone from eight 32-bit cores at 1.7 GHz

…

via Grid Blog
If you’re into this sort of thing, IBM has posted a recommended reading list for grid developers. It includes archival stuff as well as blogs (Ian Foster’s is listed, which I highly recommend).

The Register’s Ashlee Vance reports that Google takes issue with Intel’s claim that Google’s going all Xeon ga-ga. I covered the Intel side yesterday.

Ashlee actually talked to someone in the Googleplex:

“We bought a small number of chips from Intel recently, but we continue to be supplied by more than one vendor,” spokesman Barry Schnitt told us in a statement and then returned to his cocoon.

So, either: Intel’s got a loose canon who used …

Curious about how it was that a really big technical advance was announced by two separate companies at the same time, over a weekend? Me too!

The Register has a really interesting piece on how this all happened. Basically Intel is about a year ahead of IBM, but IBM got wind of Intel’s plans to announce in late January and arranged to have their hastily-patched together announcement floated at the same time as Intel’s. And, …

I’ve been following the Computing Research Association’s coverage of the impact of the government’s failure to pass a 2007 budget. Federal agencies other than DHS and DOD have been under a continuing resolution (a CR) this fiscal year. CR’s are a good thing in that they allow agencies to continue working without a budget, but they basically freeze budgets at last year’s level and don’t fund any of the new programs that were authorized.

As …

(OK. So maybe that headline is a little puerile…)

From Supercomputing Online last Friday:

The National Center for Supercomputing Applications (NCSA) today announced that State Farm, the Bloomington, Illinois-based insurer that services more than 71 million policies in the United States and Canada, is the center’s newest Private Sector Partner.