Summit (supercomputer)
Generated by Llama 3.3-70BExpansion Funnel Raw 58 → Dedup 0 → NER 0 → Enqueued 0
| Summit (supercomputer) | |
|---|---|
.jpg) | |
| Name | Summit |
| Location | Oak Ridge National Laboratory |
| Architecture | IBM Power9 CPU and NVIDIA Tesla V100 GPU |
| Operating system | Red Hat Enterprise Linux |
| Memory | 256 GB per node |
| Storage | 250 PB |
| Peak performance | 200 PFLOPS |
| Linpack performance | 143.5 PFLOPS |
| Ranking | 1st on TOP500 list |
Summit (supercomputer) is a high-performance computing system developed by IBM, in collaboration with NVIDIA, Red Hat, and Mellanox Technologies, for the United States Department of Energy's Oak Ridge National Laboratory. The system is designed to support a wide range of scientific applications, including climate modeling, materials science, and genomics, and has been used by researchers from Los Alamos National Laboratory, Lawrence Livermore National Laboratory, and Argonne National Laboratory. Summit's architecture is based on the IBM Power9 CPU and NVIDIA Tesla V100 GPU, and it has been recognized as one of the most powerful supercomputers in the world, ranking 1st on the TOP500 list, ahead of Sierra (supercomputer), Trinity (supercomputer), and Sunway TaihuLight. The system has also been used in conjunction with other supercomputers, such as Titan (supercomputer), to support large-scale scientific simulations.
● Introduction
The development of Summit was a collaborative effort between IBM, NVIDIA, Red Hat, and Mellanox Technologies, with funding from the United States Department of Energy. The system is designed to support a wide range of scientific applications, including climate modeling, materials science, and genomics, and has been used by researchers from Los Alamos National Laboratory, Lawrence Livermore National Laboratory, and Argonne National Laboratory. Summit's architecture is based on the IBM Power9 CPU and NVIDIA Tesla V100 GPU, and it has been recognized as one of the most powerful supercomputers in the world, ranking 1st on the TOP500 list, ahead of Sierra (supercomputer), Trinity (supercomputer), and Sunway TaihuLight. The system has also been used in conjunction with other supercomputers, such as Titan (supercomputer), to support large-scale scientific simulations, and has been compared to other high-performance computing systems, such as Blue Waters (supercomputer), Stampede2, and Cori (supercomputer).
● Design and Architecture
Summit's architecture is based on the IBM Power9 CPU and NVIDIA Tesla V100 GPU, with a total of 27,648 CPU cores and 27,648 GPU cores. The system has a total of 256 GB of memory per node, and a total of 250 PB of storage, provided by DataDirect Networks and Spectrum Scale. The system's interconnect is based on InfiniBand technology, provided by Mellanox Technologies, and it has a peak performance of 200 PFLOPS, making it one of the fastest supercomputers in the world, comparable to Fujitsu's K computer and Cray Inc.'s XK7. The system's design is also based on the OpenPOWER architecture, which allows for the integration of IBM Power9 CPUs with NVIDIA Tesla V100 GPUs, and has been used in conjunction with other systems, such as Google Cloud Platform and Amazon Web Services.
● Performance and Capabilities
Summit has a peak performance of 200 PFLOPS, making it one of the fastest supercomputers in the world, comparable to Fujitsu's K computer and Cray Inc.'s XK7. The system has also achieved a LINPACK performance of 143.5 PFLOPS, which is the highest performance ever recorded on the TOP500 list, surpassing Sunway TaihuLight and Tianhe-2. Summit's performance is also supported by its advanced cooling system, which is based on liquid cooling technology, provided by CoolIT Systems, and its power management system, which is based on DC power technology, provided by Eaton Corporation. The system has been used to support a wide range of scientific applications, including climate modeling, materials science, and genomics, and has been used by researchers from Los Alamos National Laboratory, Lawrence Livermore National Laboratory, and Argonne National Laboratory, as well as University of California, Berkeley, Massachusetts Institute of Technology, and Stanford University.
● Applications and Uses
Summit has been used to support a wide range of scientific applications, including climate modeling, materials science, and genomics. The system has been used to simulate the behavior of complex systems, such as weather patterns and materials properties, and has been used to analyze large datasets, such as genomic data and climate data. Summit has also been used to support research in artificial intelligence and machine learning, and has been used to develop new algorithms and models for a wide range of applications, including image recognition and natural language processing. The system has been used in conjunction with other supercomputers, such as Titan (supercomputer), to support large-scale scientific simulations, and has been compared to other high-performance computing systems, such as Blue Waters (supercomputer), Stampede2, and Cori (supercomputer).
● History and Development
The development of Summit began in 2014, when the United States Department of Energy announced plans to build a new supercomputer for the Oak Ridge National Laboratory. The system was designed and built by IBM, in collaboration with NVIDIA, Red Hat, and Mellanox Technologies, and was completed in 2018. Summit was officially dedicated on June 8, 2018, and was recognized as the world's fastest supercomputer on the TOP500 list, surpassing Sunway TaihuLight and Tianhe-2. The system has been used to support a wide range of scientific applications, including climate modeling, materials science, and genomics, and has been used by researchers from Los Alamos National Laboratory, Lawrence Livermore National Laboratory, and Argonne National Laboratory, as well as University of California, Berkeley, Massachusetts Institute of Technology, and Stanford University.
● Specifications and Configuration
Summit's specifications include 27,648 CPU cores, 27,648 GPU cores, and 256 GB of memory per node. The system has a total of 250 PB of storage, provided by DataDirect Networks and Spectrum Scale, and its interconnect is based on InfiniBand technology, provided by Mellanox Technologies. The system's peak performance is 200 PFLOPS, and its LINPACK performance is 143.5 PFLOPS. Summit's configuration includes 4,608 nodes, each with 2 IBM Power9 CPUs and 4 NVIDIA Tesla V100 GPUs, and the system is cooled using liquid cooling technology, provided by CoolIT Systems. The system's power management system is based on DC power technology, provided by Eaton Corporation, and it has been recognized as one of the most energy-efficient supercomputers in the world, comparable to Fujitsu's K computer and Cray Inc.'s XK7.