Sierra (supercomputer)

Sierra (supercomputer) is a supercomputer developed by IBM, NVIDIA, and Mellanox Technologies for the US Department of Energy's Lawrence Livermore National Laboratory. The system is designed to support the Stockpile Stewardship Program of the US Department of Energy, which aims to ensure the safety, security, and reliability of the US nuclear arsenal without conducting nuclear tests. The development of Sierra involved collaboration between Lawrence Livermore National Laboratory, Los Alamos National Laboratory, and Sandia National Laboratories. The project also received support from Oak Ridge National Laboratory and other national laboratories.

● Introduction

The Sierra supercomputer is part of the Advanced Simulation and Computing (ASC)Program of the US Department of Energy, which focuses on developing and applying high-performance computing capabilities to support the Stockpile Stewardship Program. The system is designed to provide a significant increase in computing power and data storage capacity compared to its predecessor, Sequoia. Sierra's architecture is based on the IBM POWER9 CPU and NVIDIA V100 GPU, which provides a high level of performance and energy efficiency. The system is also connected to the ESnet, a high-speed network that enables the transfer of large amounts of data between national laboratories and other research institutions, including Stanford University, MIT, and Caltech.

● History

The development of Sierra began in the early 2010s, with IBM and NVIDIA as the primary contractors. The system was designed to meet the growing demands of the Stockpile Stewardship Program, which requires advanced simulations and data analysis to ensure the safety and reliability of the US nuclear arsenal. The project involved collaboration between Lawrence Livermore National Laboratory, Los Alamos National Laboratory, and Sandia National Laboratories, as well as other national laboratories and research institutions, including UC Berkeley and Harvard University. Sierra was officially dedicated in 2018, with Rick Perry, the US Secretary of Energy at the time, attending the ceremony. The system has since been used for a variety of applications, including climate modeling and materials science research, in collaboration with NCAR and NIST.

● Architecture

Sierra's architecture is based on the IBM POWER9 CPU and NVIDIA V100 GPU, which provides a high level of performance and energy efficiency. The system consists of over 240 racks of servers, each containing multiple nodes. Each node is equipped with two IBM POWER9 CPUs and four NVIDIA V100 GPUs, providing a total of over 190,000 CPUs and 80,000 GPUs. The system also features a high-speed InfiniBand network and a large data storage system, which is connected to the ESnet and provides access to ORNL and other research institutions. The architecture of Sierra is similar to that of other supercomputers, such as Summit and Trinity, which are also used for high-performance computing applications, including research at University of Chicago and Columbia University.

● Performance

Sierra has a peak performance of over 125 petaflops, making it one of the fastest supercomputers in the world. The system is capable of performing complex simulations and data analysis tasks, including climate modeling and materials science research. Sierra's performance is also highly scalable, allowing it to be used for a wide range of applications, from small-scale simulations to large-scale data analysis tasks. The system has been used for research in collaboration with University of Michigan, UT Austin, and Georgia Tech.

● Applications

Sierra has been used for a variety of applications, including climate modeling and materials science research. The system has also been used for simulations of nuclear reactors and high-energy physics experiments, in collaboration with Fermilab and Brookhaven National Laboratory. Additionally, Sierra has been used for research in artificial intelligence and machine learning, including the development of new algorithms and models for data analysis tasks, with support from Google and Microsoft. The system has also been used for research in collaboration with UCLA, UIUC, and Carnegie Mellon University.

● Upgrades_and_Development

Sierra is expected to undergo upgrades and development in the coming years, including the addition of new hardware and software components. The system is also expected to be used for new applications, including quantum computing and cryptography research, in collaboration with IBM Quantum and NSA. The development of Sierra has also led to the creation of new technologies and innovations, including the development of new materials and manufacturing techniques, with support from DARPA and NSF. The system has also been used for research in collaboration with UW-Madison, UW, and Duke University. Category:Supercomputers