I was featured fairly prominently in a recent CNET article titled “Help! I can’t program my car.” The story centered on how new cars are are becoming increasingly computerized and that snazzy new features now require a significant amount of…

One of the stops on the press preview of the exhibit hall was the RapidMind booth, where we learned of some astonishing performance increases made possible by their Multi-Core Development Platform. The platform is comprised of a Code Optimizer (analyzes and optimizes computations), Load Balancer (synchs work to keep all cores fully utilized), Data Manager (reduces bottlenecks), and Diagnostics. The API integrates with C++ Windows or Linux apps and requires no new tools, compilers, or workflow. Applications that are not multi-core enabled can take advantage of all available cores (including GPUs and Cell processors) and will automatically scale to future additional cores.

Just when I finally convinced myself to get off of the couch/ away from the computer screen and actually you know DO something, I learned about ResearchChannel. This nonprofit media and technology organization makes the work of researchers available to…

Our noble goal of visiting every display at SC07 devolved, due to the crowds and number of booths, to passing slowly by every display and stopping to chat when we could. Two spots, however, commanded daily pilgrimages: NASA and Fusion-io….

Few displays here on the exhibit hall floor in Reno could create more slack-jawed admiration than that of HLRS. We gathered around the big screen to watch the collaborative, real-time viewing and manipulation of a 3-D car engine model by…

Among the many impressive stories to be heard at the Ohio Supercomputer Center booth here in Reno is that of its Blue Collar Computing initiative. And how it saved Big Bird. But first things first

Blue Collar Computing was launched in 2004 to give small- and medium-sized companies access to the kind of virtual modeling and simulation that would ordinarily be available only to Fortune 500-type behemoths. They describe their efforts as “democratizing supercomputing” by providing hardware, software, and training that would be out of reach to us mere mortals who can’t put a high-performance server on the VISA card, spend thousands on software licenses, and hire a squadron of programmers.

At GCG, we’re all about business value. As much as we love the new new thing, it has to pencil out. That said, we sat in on IBM’s “Mainframe Gas Gauge” teleconference to learn more about monitoring power usage and…

IBM Corporation announced today that a new process for reclaiming silicon wafers has been developed at its Burlington, Vermont facility.  Until now, silicon wafers have been scrapped at a rate of 7,500 per day because they contain intellectually property and…

We’ve been a bit lazy on the blogging front, a bunch of business travel, coupled with the demands of getting this particular survey out the door have pushed blogging down a couple of pegs on ‘gotta get this done now’ list.

This new survey has taken up quite a bit of bandwidth around here. We had 297 enterprise x86 server customers answer a whole bunch of questions including: they’re dealing with power/cooling/floor space issues, which vendor does the best job of quoting/pricing deals, what they’re doing about x86 virtualization, which chip vendor has captured their heart (Intel, AMD, neither?), along with all of the vendor vs. vendor ratings on technical (performance, availability, quality, etc.), customer support (service, promise keeping, advice, etc.), and other issues (road maps, R&D prowess, trustworthiness, etc.)