NVIDIA is among six technology companies to receive funding from the U.S. Department of Energy’s Exascale Computing Project (ECP) to accelerate the development of next-generation supercomputers.
The ECP mission is to facilitate the delivery of at least two exascale computing systems, with an aim to deliver at least one by 2021. Such systems would be approximately 50x more powerful than the nation’s fastest supercomputer, Titan, located at Oak Ridge National Laboratory, in use today.
The goal of the ECP PathForward program is to find solutions that maximize the energy efficiency and overall performance of future large-scale supercomputers critical to areas such as national security, manufacturing, industrial competitiveness and energy research.
In addition to performance, the DOE has ambitious goals for improving power efficiency, to achieve exascale performance using only 20-30 megawatts. By comparison, an exascale system built with CPUs alone could consume hundreds of megawatts.
NVIDIA has been researching and developing faster, more efficient GPUs for high performance computing for more than a decade. This is our sixth DOE research and development subcontract, which will help accelerate our efforts to develop highly efficient throughput computing technologies to ensure U.S. leadership in HPC.
Our R&D will focus on critical areas including energy-efficient GPU architectures and resilience. Our findings may be incorporated into future generation GPU architectures after Volta (which will be used in the DOE’s upcoming flagship Summit and Sierra supercomputers, scheduled to go online in 2018).
The DOE has placed a high priority on supercomputer research. Its PathForward technical requirements state, “The U.S. faces serious and urgent economic, environmental, and national security challenges based on energy, climate, and growing security threats. High performance computing is a requirement for addressing such challenges, and the need for the development of capable exascale computers has become critical for solving these problems.”
To facilitate and test our technology, NVIDIA research teams will collaborate closely with six national DOE laboratories: Argonne National Laboratory, Lawrence Berkeley National Laboratory, Lawrence Livermore National Laboratory, Los Alamos National Laboratory, Oak Ridge National Laboratory, and Sandia National Laboratories.