It’s been more than a decade since IBM’s supercomputer at the heart of the company’s Data Center Security program went online.
Today, the company is making another big step toward that goal.
The supercomputer is part of the giant $6.4 billion project to build a 1,200-petaflop supercomputer in Arizona, which is one of the nation’s largest supercomputers, and one of only a handful of such systems built by IBM in the country.
“I can’t think of a single company that has a larger computing machine in the world that has not had a major supercomputer project underway for a while now,” IBM CTO Bob Jacobs said at the company conference this week.
In the coming months, the supercomputer will be built on the companys $6 billion, 740-peta supercomputer farm in Arizona and in collaboration with the Arizona Department of Energy.
It will also be the first IBM supercomputer built in Arizona.
In its current form, the Arizona project is a massive undertaking that requires the company to develop a new generation of supercomputing hardware and software that will enable it to compete with other big companies in the data center market.
The project has been in the works for a few years.
“The Arizona project, while not a huge one, has become very large in the past few years,” Jacobs said.
“It’s not just about building a new supercomputer.
It’s also about bringing that new super computer to the Arizona market and providing it to our customers, to our partners, and to the general public.”
The state-of-the-art supercomputer includes all the features that IBM expects will be required to compete in the supercomputation market, including: 1,000 cores, or 1,024 processors, that are capable of building large data sets in real time, with each core processing the same amount of data.
“In a data center, a data processor can make use of multiple cores to do different tasks, and in a supercomputer, you’re talking about processing a very large amount of work simultaneously,” Jacobs explained.
“This means that the cores are working in parallel, so there’s no limit on the number of cores you can put together.”
The supercomputed data is then fed into an array of servers and the super computers, which can process it on the same hardware as a regular server.
The data is processed by a large number of computers running at various speeds.
IBM has said the super computing technology will be “relatively low-power.”
“We’re talking tens of teraflops [trillion] per day.
The amount of processing is extremely low,” Jacobs told the audience at IBM’s annual shareholder meeting in May.
IBM also announced the creation of a new research and development lab in Arizona that will develop and commercialize new and innovative technologies for the super computation industry.
The Arizona project has also been a key component of IBM’s effort to compete more closely with its fellow supercomposers, which include Amazon Web Services and Microsoft Azure.
The company said the new supercomputer is expected to support a variety of industries, including manufacturing, energy, financial services, healthcare, education, and telecommunications.
IBM said the Arizona program will also support “several major research projects, including a multi-year collaboration with a leading provider of deep learning systems for the energy sector.”
IBM said it expects the Arizona supercomputer to be ready by 2022.
“By 2021, we expect to have a supercompleter system ready to compete on a par with existing supercomprehensive systems, and will be able to offer it to a variety in-demand customers,” Jacobs added.
“And, by the end of that year, we believe that we will be capable of providing a comparable supercompletion to a super server in other parts of the globe.”
IBM is building a $10 billion supercomputer near the Utah border, called Alpha, with IBM Research and Microsoft, with plans to complete the super computer by 2025.
The other Arizona supercompletion is expected later this year.
The new super computers will be powered by IBM’s POWER5 processor, which will be used to process applications that are larger than 1 gigabyte in size, such as database queries.
IBM is planning to use the super-efficient Power5 to accelerate some of the more advanced applications in the company�s Power Cloud platform, including data analytics.
The POWER5 will also power the new Arizona super computer and a new POWER7 supercomputer system.
“We�re excited about the progress we�ve made in our partnership with the government to build the Arizona Project, and look forward to working with the industry as we prepare for the next generation of IBM supercomposition,” Jacobs continued.
“As the world moves toward more cloud computing, and the demand for supercompsets increases, the industry is looking for new and different ways to combine our