January 11, 2012
By Anthony J. Lockwood
Dear Desktop Engineering Reader:
With the advent of powerful workstations, these are goods times for practitioners of compute-intensive life science applications such as bioinformatics, molecular dynamics, and quantum chemistry. The times may just have gotten a lot better. But don’t take my word for it. Benchmark one of your models and see for yourself.
Microway recently announced SimCluster, which it describes as an off-the-shelf, integrated cluster designed for life science researchers. SimCluster is said to be “fully” optimized for simulating large models and achieving higher accuracy. It comes with NVIDIA Tesla M2090 GPUs and Intel Xeon 5600 series CPUs connected by InfiniBand, Linux or Windows, NVIDIA Tesla Bio Workbench applications, the NVIDIA CUDA parallel programming environment, and cluster management tools.
Really good stuff right there, but not the big cigar in this announcement. What I like about Microway’s approach to their SimCluster announcement is that they put their cards right on the table: They offer a “try before you buy” program. And your participation in it is a no-brainer that you can handle from your desk.
How the program works is simple: You sign up to gain access to a four-node, eight GPU SimCluster. You then upload one of your AMBER, MATLAB, or NAMD models to a SimCluster and run it. (You can upload your custom CUDA code too, but that requires a little advanced planning.) This lets you benchmark your present performance against the SimCluster’s speed. If you like what you see, Microway says that they can ship you the exact configuration within 48 hours (custom-specked SimCluster configurations take longer, of course).
Microway’s goal here is to build a path for research groups to move from GPU-accelerated workstations to a SimCluster configuration so that they can accelerate their research. That’s not to say they are leaving the GPU-accelerated workstation world behind. The company offers its WhisperStation line of 64-bit Linux or Windows workstations for users with smaller budgets and less-demanding application requirements. One version in particular that’s said to be suitable for life science applications offers four NVIDIA Tesla C2075 GPUs—no slouch there.
Be that as it may, the SimCluster “try before you buy” program seems to me to be a golden opportunity for you to see just what stepping up to a cluster can mean for you. Pick something gnarly and check it out. You’ve nothing to lose, but maybe a lot of productivity to gain. You can learn more about the SimCluster and the benchmark program from the links at the end of today’s Pick of the Week write-up.
Anthony J. Lockwood
Editor at Large, Desktop Engineering
This is sponsored content. See how it works.
About the AuthorAnthony J. Lockwood
Anthony J. Lockwood is Digital Engineering’s founding editor. He is now retired. Contact him via [email protected].Follow DE