Andy Keane, the general manager of our Tesla division, has a great op-ed piece in All Things Digital this week, warning that America’s leadership in supercomputing is at risk.
These days, everything from new shampoo to energy exploration depends on supercomputers. But, as he writes in “Crisis in Computing,” U.S. supercomputing resources are in surprisingly short supply, with most research supercomputers 2x oversubscribed. Without a concerted investment around a national high performance computing (HPC) infrastructure, we’ll lose ground to other nations – with resulting losses in business growth, job creation and competitiveness.
Ironically, it’s the U.S. history of leadership in supercomputing that has put us in this vulnerable spot. Our traditional CPU-based technology infrastructure is no longer able to meet our computational needs in a way that’s cost or power efficient. Meanwhile, other nations are free to leapfrog past legacy systems and go directly to HPC architectures using both GPUs and CPUs for increased efficiency and performance at lower prices.
Computational demand may increase 1000x in the next 10 years. Will we maintain our stature in HPC or watch our advantages slip away to other players? Check out Andy’s article and let us know what you think.
Tagged: Supercomputing, Tesla
