The U.S. Department of Energy and the Lawrence Livermore National Laboratory (LLNL) last week announced the unveiling of Sierra, one of the world’s fastest supercomputers. Sierra is ranked as the world’s third-fastest supercomputer in the world with a peak performance of 125 petaFLOPS -125 quadrillion floating-point operations per second.
“Sierra can perform most required calculations far more efficiently in terms of cost and power consumption than systems consisting of CPUs alone,” the LLNL team said in a press release.
The system is comprised with over 17,000 NVIDIA Tesla V100 GPUs connected via NVLink interconnect.
“The system’s NVIDIA GPUs also present scientists with an opportunity to investigate the use of machine learning and deep learning to accelerate time-to-solution of physics codes,” the LLNL team stated. “It is expected that simulation, leveraged by acceleration coming from the use of artificial intelligence technology, will be increasingly employed over the coming decade.”
The computer has a footprint of 7,000 square feet and includes 240 computing racks and 4,320 nodes.
“Sierra is a world-class, pre-exascale supercomputer that allows researchers to run large complex scientific simulations at scale, at speeds never before thought possible,” said Ian Buck, vice president and general manager of Accelerated Computing at NVIDIA.
The supercomputer was designed and built by IBM.
Read more>
Lawrence Livermore Unveils Sierra, World’s Third Fastest Supercomputer
Oct 31, 2018
Discuss (0)
Related resources
- GTC session: Accelerate your Workflows and Gain Competitive Advantage With AI Workstations (Presented by Z by HP)
- GTC session: Live from GTC: A Conversation on the Latest in HPC
- GTC session: Which GPU for Open Source LLM Applications? Best Practices on Hyperstack (Presented by NexGen Cloud)
- NGC Containers: CUDA
- NGC Containers: Animation Graph Microservice
- SDK: cuSOLVERMp