At SC09, it was hard to travel the floor without running into hybrid computing in some way, shape, or form. In fact, you couldn’t swing a cat without hitting someone or something related to using accelerators to maximize performance. (I know this because I actually did bring in a cat to test this theory… an actual live cat… well, it was alive at the beginning of the testing.) The chief beneficiary of all of this interest in hybrid computing looks to be NVIDIA who, with their Tesla line of GPU accelerators, is currently offering the most robust product line.

‘Read More’ to see our video interview with NVIDIA…

As we noted in two previous blogs, large shared-memory computing seems to be making a bit of a comeback. The two players we talked about earlier, ScaleMP and 3Leaf, use primarily off-the-shelf servers (with a special ASIC, in 3Leaf’s case) along with Infiniband connections and specialized software to build cache-coherent, shared-everything systems. With this, they can build systems spanning 16 motherboards – which is a pretty damned big system when you factor in the core counts you can get today. SGI is taking a different, and much larger, swing at this with their new UltraViolet Nehalem-based clusters.

‘Read More’ to see our video interview with SGI…

OK, sure: liquid can hold and transfer way, way more heat than air – we all know that. But is dropping an entire rack of servers into what looks like an enormous deep fryer the right solution? The answer from the folks at Green Revolution Cooling is a resounding, “Yes!”

‘Read More’ to see our video interview with Green Revolution…

It’s not too hard for GCG staff to find smart technology people to explain stuff to us. It is, however, hard to find someone this smart – meet Dave Turek, head of IBM’s Deep Computing division. In this interview available…

As we talked about in our last blog (or in a different blog, if you’re not reading these in order), a handful of vendors are pushing technology that lets users tie bunches of smaller systems into large, shared-everything SMP servers. ScaleMP was founded in 2003 and has been shipping products since 2006. They are pitching a software-based solution that allows customers to combine up to 16 x86 nodes into a large, single-image system. This can give customers up to 128 cores and 4TB of RAM in a single o/s instance to use with apps that need huge memory or can’t be easily parallelized. However, as the company points out, even highly parallelized message-passing apps perform faster on shared memory systems.

‘Read More’ to see our video interview with ScaleMP…

The wheel is turning full circle. In the 1980s and 1990s, the HPC crowd pioneered scale-out processing and advanced parallelism in response to cost and scalability limitations imposed by the shared-everything systems of the day, ushering in an era where huge numbers of small computers were harnessed to solve large-scale problems.

‘Read More’ to see our video interview with 3Leaf…

As we’re slowly gathering our wits again after a week of supercomputing fun at SC09, a story from Chris Mellor caught our eye this morning. Chris was writing about Fusion-io and their newest SSD (Solid State Drive) product, the Fusion ioDrive Octal. These units provide up to 5TB raw storage and can push up to 800,000 IOPS.

‘Read More’ to see our video interview with Fusion-io…