By Doug Barney
A few newsletters ago I complained that Microsoft Corp. of Redmond, WA. (www.microsoft.com) wasn’t doing enough on the client side to exploit multicores, an opinion seconded by the folks from RapidMind.
The same isn’t quite so true for servers. On the commercial side, Microsoft has had Compute Cluster Server 2003 for four or so years, and is now wrapping up the upcoming HPC Server 2008.
The most impressive HPC area is where Microsoft Research works with the world’s top scientists on data mining, handling massive data sets, and applying HPC techniques to solving the earth’s biggest problems. (Here and here are two articles I’ve written about these efforts.
To further those efforts, Microsoft added to its HPC/supercomputing brain trust by hiring Daniel A. Reed, Microsoft’s new director of scalable and multicore computing. Reed’s latest position was as the director of the Renaissance Computing Institute in North Carolina.