In a press release on Monday, NVIDIA Longre announced that the Los Alamos National Laboratory (LANL) "Venado" supercomputer driven by grace GPU + hopper supercomputer will have an incredible 10 exaflops peak AI performance** It is reported that a series of world leading computer manufacturers, including ATOS, Dell, gigabyte, Huiyu, Langchao, Lenovo and supermicro, are using grace supercomputers to build the next generation of servers.
(from: NVIDIA newsroom ) The
As the supercomputing field is moving towards billions, NVIDIA grace superchips will be able to provide support for the new generation of artificial intelligence (AI) and high-performance computing (HPC) workloads.
All of these new systems benefit from the grace and grace hopper design in the just announced HGX platform, which provides manufacturers with the blueprint they need to build the system, supplemented by leading data center CPU performance and twice the memory bandwidth / energy efficiency.
Ian buck, vice president of NVIDIA ultra large scale and high performance computing, said:
As supercomputing has entered the AI era of billions of times, NVIDIA is working with OEM partners to help researchers cope with the great challenges that were far away in the past.
No matter in the fields of climate science, energy research, space exploration, digital biology, or quantum computing, grace CPU and grace hopper super chip can form the basis of the most advanced HPC / AI platform in the industry.
Among the first batch of systems confirmed to adopt NVIDIA grace super chips, customers in the United States and Europe are included. For example, Los Alamos National Laboratory (LANL) announced on the same day that its next generation Venado supercomputing system will take the lead in adopting NVIDIA grace CPU technology.
It is reported that Venado is built based on HPE Cray ex heterogeneous supercomputing system, which mixes grace CPU and grace hopper superchip nodes, and is applicable to a wide range of emerging application sets. Upon completion, its estimated AI performance will exceed 10 exaflops.
Irene qualters, deputy director of simulation and computing at LANL, said:
With the performance of NVIDIA grace hopper, LANL researchers can continue to work to push the boundaries of scientific breakthrough through Venado.
NVIDIA's accelerated computing platform and broad ecosystem are working to remove performance barriers, so that LANL can make new discoveries to benefit the country and the whole society.
As for the new system of the Swiss National Computing Center, Alps also chose to build it based on HPE Cray ex supercomputing. As a general-purpose system open to the research community in Switzerland and other parts of the world, it also hopes to achieve new breakthroughs in a wide range of research fields with the help of grace CPU super chip.