My associate Dr. Otero commented on his blog “Like it or not, life science researchers at large need to recognize that their research is intertwined with supercomputing resources in one way or another.”

I’d take that one step further than just life sciences and say that most science and engineering research is or is going to be intertwined with supercomputing. As stated in a 2008 TOP500.org article by Hans Werner Meuer, “… it will take six to eight years for any system to move from position one to [position] 500 [on the Top500 list].”

With a Petaflops system entering the list in 2008, this means that today’s researchers will have ample access to Petaflops systems in the coming years. What this means at Dell is that we are thinking now in terms of Petaflops of compute power. And the software, interconnects, storage, file systems and service-and-support to handle that computer power. We have a short six years of runway before Petaflops will become commonplace. Dell has in place a new HPC Team with management support with this as our focus.

With new GPGPU products to be available this year, a mere 8 racks will have a Petaflops peak rating. Projections deliver Petaflops within a single rack around 2017. The affects on numerous types of scientific and engineering research are unimaginable.

Dell is poised to deliver these commodity Petaflops and empowering such scientific and engineering. We remain focused on efficient, commodity HPC, which is quickly growing to be Petaflops. Commodity Petaflops enable Exaflops.

To echo but paraphrase Dr. Otero, the future of the human race depends on HPC. Just some thoughts as we enter the next decade.

-- Mark R. Fernandez, Ph.D.