Alissa Tabirian
NS&D Monitor
10/23/2015
The Lawrence Livermore National Laboratory (LLNL) has awarded a subcontract to Silicon Valley-based Penguin Computing for stockpile stewardship computing services at three National Nuclear Security Administration (NNSA) laboratories, the NNSA announced this week. The contract supporting the NNSA’s Advanced Simulation and Computing (ASC) program involves a $39 million award to the company for “7 Petaflops (quadrillion floating operations per second) of ‘capacity’ computing capability to Los Alamos, Sandia and Lawrence Livermore national laboratories,” according to the announcement. As a result, it says, NNSA supercomputers will be “dedicated to the largest and most complex calculations critical to stockpile stewardship,” the program that maintains the U.S. nuclear deterrent without full-scale explosive testing.
The NNSA said each lab will receive computing clusters between 2016 and 2018 with scalable units, each of which “represents approximately 200 teraflops of computing power” and can be connected to other units “to create more powerful systems.” The announcement notes that the “tri-lab procurement model reduces costs through economies of scale based on standardized hardware and software environments at the three labs,” adding that advances in computation technology in the last two decades have led to major reductions – by a factor of 20,000 – in the cost of high-power computing systems.
The ASC program develops computer simulation capabilities “to analyze and predict the performance, safety, and reliability of nuclear weapons and to certify their functional,” according to the NNSA. An event held this week to mark the 20th anniversary of the Stockpile Stewardship Program highlighted one of the program’s most difficult endeavors: “achieving the high performance computing systems required to simulate the nuclear explosive process at high fidelity.” The NNSA said that as a result of two decades of work, “a whole new computational model, based upon massively parallel processing, has now become a global high performance computing standard.”
