First USGS Supercomputer Decommissioning makes way for Successors
In 2013, the U.S. Geological Survey welcomed its first general-use supercomputer available to all research staff.
Supercomputers, also called a High-Performance Computing, or HPC, systems, are powerful machines capable of quickly working on large-scale computations. They allow scientists to gather and analyze data more quickly and simulate complex experiments that may be too impractical to conduct on smaller machines in reasonable amounts of time.
Prior to the availability of supercomputers at USGS, scientists built their own by networking together various components into compute clusters, sometimes referred to as a “Frankenstein machine”. One such machine in the newly created Science Analytics and Synthesis group demonstrated the need to invest in larger scale High Performance Computing systems for the acceleration and expansion of scientific discovery. This ultimately led to the purchase of the very first supercomputer and creation of the Advanced Research Computing office.
This HPC, named Yeti, was installed at the Bureau of Land Management Data Center at the Denver Federal Center. The name Yeti was a way to honor the contributions of the early “Frankenstein machine” predecessors.
Yeti had 3,728 central processing unit (CPU) cores and 35,072 CUDA GPU cores, with 32,128 GB of random-access memory (RAM). To put this into perspective, a typical laptop computer usually has 4 cores and 8 GB RAM.
A decade of growing computer power
Over the years, the Advanced Research Computing team has grown to a staff of eight, and now hosts over 20 user trainings annually for any interested USGS scientists or research staff member to come and learn more about the application of using supercomputers as a tool in their research.
The trainings are designed to help get researchers gain the experience necessary to begin working on HPC resources as soon as possible. To facilitate this, a broad course offering is available including introductory offerings such as HPC 101 to advanced topics like Artificial Intelligence (AI)/Deep Learning.
The staff also provide weekly office hours to allow users the opportunity to ask specific questions and gain resources to troubleshoot any issues they are having concerning their work on the HPC.
These trainings and technical support sessions have been very well received by scientists, who rely on ARC staff to assist when researchers use the HPC machines. During Yeti’s tenure it had been the primary machine utilized by staff during training sessions.
By the numbers
In all, Yeti helped facilitate a vast amount of work including over 6.2 million computing jobs. It has assisted over 1038 users and was used to train over 600 users, including 75 users from federal agencies outside of the USGS.
“To help illustrate the significance of Yeti’s work output over 5 years, one could imagine that same workload being run on a single CPU, would have taken approximately 14,500 years by comparison,” said Jeff Falgout, technology lead for the ARC team.
On August 16, 2023, a ribbon cutting took place celebrating the installation of Yeti’s replacement and newest USGS supercomputer addition, this time at the Earth Resources Observation and Science Center (EROS) in Sioux Falls, SD. Named Hovenweep after the national monument along the Utah-Colorado border, this machine is currently the most powerful computer in the Department of the Interior. It joins supercomputers Tallgrass and Denali already in operation as the next generation of high performance-computing at USGS.
Hovenweep is the third HPC system to be installed at EROS, where it will continue to advance scientific understanding and management of environmental events such as wildfire risk to homes, extreme weather events such as emergency flood mapping, and volcano, landslide, and tsunami hazards.
Tallgrass, named after the Tallgrass Prairie National Preserve in Kansas, was added in 2020. It was designed for AI and analytics workloads. Tallgrass is equipped with both the hardware and software to handle modern AI, deep learning, and machine learning demands.
Denali is named for the Denali National Park & Preserve in Alaska and was operational at the EROS in Sioux Falls, SD by January 15, 2020. Denali was designed to be used for large scale simulations and analytics.
“It is important to note that future growth of the USGS HPC program is in part, reliant on researchers citing or acknowledging these resources in their publications to show their value to their work.” said Janice Gordon, lead of Advanced Research Computing team. “The USGS supercomputers have become indispensable tools for researchers and their importance only seems to be growing as an asset for future science.”
Get Our News
These items are in the RSS feed format (Really Simple Syndication) based on categories such as topics, locations, and more. You can install and RSS reader browser extension, software, or use a third-party service to receive immediate news updates depending on the feed that you have added. If you click the feed links below, they may look strange because they are simply XML code. An RSS reader can easily read this code and push out a notification to you when something new is posted to our site.