The National Center for Atmospheric Research (NCAR) will install a new 5.34 petaflops supercomputer – Cheyenne – from SGI and a 20PB high performance storage system from DataDirect Networks, both companies announced today. Cheyenne, an SGI ICE XA system, is expected to be operational in early 2017 and will be powered by more than 7000 ‘future generation’ Intel processors.
Cheyenne will be roughly twice as powerful as NCAR’s existing system, Yellowstone, and is intended to perform some of the world’s most data intensive calculations for weather and climate modeling to improve the resolution and precision by orders of magnitude.
“NCAR requires an increasingly more advanced system. For example, doubling the resolution of a weather system requires a tenfold increase of compute power. This is necessary to help model and pinpoint when and where a storm will hit. To put it simply, we’re aiming to process data of very large proportions,” said Al Kellie, director of Computational & Information Systems Lab at NCAR.”
“Having a centralized, large-scale storage resource delivers a real benefit to our scientists,” said Anke Kamrath, director of the operations and services division at NCAR’s computing lab. “With the new system, we have a balance between capacity and performance so that researchers will be able to start looking at the model output immediately without having to move data around. Now, they’ll be able to get right down to the work of analyzing the results and figuring out what the models reveal.”
Sponsored by the National Science Foundation, NCAR focuses on understanding the Earth’s atmosphere and related geospace systems. NCAR scientists will use the SGI system to help forecasters generate more detailed alerts of impending solar-induced geomagnetic storms in the Earth’s upper atmosphere.
Pairing the SGI machine with DDN’s new SFA14K high-performance hyper-converged storage platform should enable NCAR to perform sophisticated modeling while substantially reducing workflow bottlenecks. As a result, the organization will be able to quickly process mixed I/O workloads while sharing up to 20 PBs of vital research data with a growing scientific community around the world.
After implementing the Cheyenne supercomputer, NCAR scientists expect to better understand how particular regions across the globe will be impacted by rising sea levels and changing patterns of storms, precipitation, and temperature. The new supercomputer will also be used to predict climate patterns over the next ten years or further into the future to assess drought risk or the extent of melting sea ice in the Arctic.
In addition, NCAR research will lead to improved predictions of severe atmospheric events, such as hour-by-hour risks associated with thunderstorm outbreaks to the timing of the 11-year solar cycle and its potential impacts on GPS and other sensitive technologies. Cheyenne is expected to be a critical tool for researchers across the country studying seismic activity, air quality, wildfires, and other important geoscience topics.
“NCAR’s new supercomputer, Cheyenne, will be an important tool for researchers across the country to understand climate change, severe weather, air quality and other important atmospheric and geoscience topics,” said Jorge Titinger, president and CEO, SGI.
Snapshot of key system attributes:
- Three times more energy efficient in flops per watt than Yellowstone.
- Has a modular design enabling combinations of processors and accelerators
- Includes more than 7000 future generation Intel Xeon processors in high density packaging with SGI E-cell warm water cooled technology for lower power consumption
- Features a high-performance enhanced hypercube interconnect based on Mellanox EDR InfiniBand and 20 Petabytes of High Performance Storage from Data Direct Networks (DDN)
- Incorporates SGI HPC Software including SGI Performance Suite to deliver parallel application performance and SGI Management Suite to enable auto-configurability, power management, health management, and remote system management
- Is capable of more than 2.5 times the amount of scientific computing performed by Yellowstone, the current NCAR supercomputer, and interoperates with existing file systems. (For more information, please visit: sgi.com/weather.)
“The new system will enable us to keep pace with the increased number of people trying to do very large data assimilation problems,” added Rich Loft, director of technology development in the computational and information systems laboratory at NCAR. “Earth system research is very data intensive. NCAR will now be able to do more to help scientists go beyond just studying phenomena to actually making predictions through data-intensive simulations that require larger I/O bandwidth and storage performance.”
“Weather and climate simulations require moving and analyzing increasingly large amounts of data. Mellanox InfiniBand solutions accelerate the overall system performance by enabling in-network computing, the ability to analyze data as it moves within the datacenter, which reduces simulation time,” said Marc Sultzbaugh, senior vice president at Mellanox Technologies.
NCAR also hopes to leverage DDN experience gained from broad global deployments in weather and climate modeling to support data-intensive research efforts such as the Coupled Model Intercomparison Project phase 6 (CMIP6). CMIP6 is a multi-model experiment framework and involves the execution and analysis of high-resolution, interlocked climate models and global experiments. NCAR’s will also leverage DDN’s long-standing technology partnership with SGI to optimize traditional climate codes, including the Weather Research and Forecasting Model (WRF) and Community Earth System Model (CESM).