NorthIowaToday.com

Founded in 2010

News & Entertainment for Mason City, Clear Lake & the Entire North Iowa Region

Iowa State turns on ‘Cyence,’ the most powerful computer ever on campus

Arun Somani led the team of researchers that brought Cyence to campus. Larger photo. Photo by Bob Elbert.
Arun Somani led the team of researchers that brought Cyence to campus. Larger photo. Photo by Bob Elbert.

AMES, Iowa – The most powerful computer ever on the Iowa State University campus – a machine dubbed “Cyence” that’s capable of 183.043 trillion calculations per second with total memory of 38.4 trillion bytes – is just beginning to produce data for 17 research projects.

The thinking and infrastructure behind the new machine will have much broader effects across the university.

“The whole campus is excited about this and so am I,” said Arun Somani, an Anson Marston Distinguished Professor in electrical and computer engineering and associate dean for research in the College of Engineering. “We expect to expand our science research with the help of high performance computing. We also expect this will expand Iowa State research. If you don’t have a machine powerful enough to do the calculations, you can’t even propose a project.”

Cyence, with its 12 black cabinets and rows of blue lights, is front and center in the Machine Room in the basement of the Durham Center. It has been running since early July. It succeeds Cystorm and its 28.16 trillion calculations per second as the most powerful computer on campus.

It was purchased with the help of a three-year, $1.8 million major research instrumentation grant from the National Science Foundation. Another $800,000 is being provided by Iowa State’s Office of the Vice President for Research and Economic Development and the colleges of agricultural and life sciences, engineering and liberal arts and sciences.

Somani led the faculty team that applied for the National Science Foundation grant. Other team members include Rodney Fox, Anson Marston Distinguished Professor in chemical and biological engineering and affiliate of the U.S. Department of Energy’s Ames Laboratory; Mark Gordon, Distinguished Professor in chemistry and affiliate of the Ames Laboratory; Gene Takle, professor of agronomy and geological and atmospheric sciences; and Srinivas Aluru, formerly an Iowa State professor of electrical and computer engineering now at Georgia Tech in Atlanta.

The team proposed that Cyence be used to support 17 research projects from eight Iowa State departments, including work in bioscience, chemistry, ecology, fluid dynamics, atmospheric science, materials science and energy systems.

Somani, for example, will use the new computer to help develop new and better infrastructure designs for the country’s energy and transportation systems.

“This computer enables us to solve larger models covering longer time frames,” Somani said. “We can do a lot more. This will help us a lot.”

Iowa State’s new approach to high performance computing will also make a difference across campus.

Cyence can do 183 trillion calculations per second. Larger photo. Photo by Bob Elbert.
Cyence can do 183 trillion calculations per second. Larger photo. Photo by Bob Elbert.

Under the new model, the university (with support from the Vice President for Research and Economic Development, Information Technology Services plus the colleges of agriculture and life sciences, engineering and liberal arts and sciences) will cover the infrastructure and operating costs for two high performance computing machines. New machines will be acquired every five years or so.

One machine, as was the case with Cyence, will be supported by grants and external funding. The second will be supported by researchers from across campus pooling resources to build and share a machine. The second approach is called a “condo cluster.”

“The idea behind this is that faculty develop common requirements, pool their funds and buy a much larger system than they could individually,” said Jim Davis, Iowa State’s vice provost for information technology and chief information officer. “Costs are kept low by sharing support and infrastructure, and by pooling unused capacity from all stakeholders. Researchers can run jobs and simulations that are much larger than they would be able to otherwise.”

Somani said researchers are already beginning to pool resources for a second high performance computer. He expects that machine to be up and running by next July 1.

And while Cyence isn’t going to be fast enough to make the list of the world’s top supercomputers, it is going to make a difference on campus, Davis said.

“It will make an impact on science,” he said. “We are providing facilities that faculty can use to accelerate their research work.”

0 0 votes
Article Rating
Subscribe
Notify of
guest

0 Comments
Inline Feedbacks
View all comments

Even more news:

Watercooler
Copyright 2024 – Internet Marketing Pros. of Iowa, Inc.
0
Would love your thoughts, please comment.x
()
x