Saturday, May 18, 2024

Sun, IBM Forge High Performance Computing Units

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

Sun Microsystems and IBM are once again going head-to-head in their efforts to lure new customers. Both companies have announced plans to reshape their high-performance computing offerings through the formation of new business units.Sun earlier this month announced the creation of its High Performance and Technical Computing (HPTC) unit and IBM this week unveiled its Deep Thinking unit. Both firms want use these ventures to rein in their intensive computing efforts by better defining them to serve customers. They believe creating HPTC systems that draw resources from their software, services, sales and storage divisions should help them do that.

HPTC is a market segment that uses complex, large-scale computers to perform any number of tasks from gauging weather patterns to running tests in life sciences organizations. Grid computing — the application of many computers’ resources to a single problem at the same time — is one facet of the HPTC segment. Until now, both Sun and IBM have been rolling out separate computer systems piecemeal, but by integrating them with their other hardware and software product offerings, they hope they will be better able to market these high-priced systems to prospective customers in need of serious number crunching.Better visibility, they hope, will lead to bigger contracts with educational institutions, government offices, and even financial organizations, in a time when competition is cutthroat and other competitors such as HP, SGI and Cray are hungry for pieces of the multi-billion-dollar HPTC pie.

Ed Broderick, principal analyst at the Robert Frances Group, said the move was an obvious one, and that the close timing of the announcements is an example of the “me, too, me, too” competitive stance the firms have struck up over the years.

“They are focusing on a new opportunity, a new initiative to bring their skills and power into one organization for HPTC,” Broderick told “This is for synergy and they’re saying ‘let’s exploit it.'”

Sun discussed its HPTC plans in a call with reporters and analysts Wednesday. Clark Masters, Executive Vice President of Sun’s Enterprise Systems Products and Shahin Khan, newly-anointed Vice President of the HPTC unit, highlighted where they see areas of growth and expressed confidence that HPTC will thrive despite some evidence that low-level Linux clusters are eating away at the high-performance market.

Masters, acknowledging the trend of low-level Linux clusters eating at HPTC market share, said Sun isn’t concerned because it has been engaged in contract talks with several government organizations, especially at the intelligence level, who indicated interest in HPTC systems.

Masters said research and development is underway to get the Java software language up to high-performance computing speed, something it is not currently equipped to handle. Sun’s labs are experimenting with this, among other HPTC challenges, and Masters said Java Grandewas created to use “Java as more than just a wrapper.” He also advised the public to expect an “aggressive and complete Linux offering soon,” but didn’t stray farther than that.As for competition, Masters entertained a question about how IBM’s Regatta stacks up against Sun machines in the HPTC department, pointing out that Sun has better input/output bandwidth and processing, while IBM enjoys better floating points per CPU to Sun’s machine.

Masters also said IBM paid Sun a compliment by ratcheting up their own HPTC interests as Big Blue this week announced the creation of the Deep Computing business unit, which, like Sun’s HPTC, is intended to link together the company’s hardware, software and services offerings for intensive computing projects.

This group will be led by Dave Turek, who most recently focused on grid computing and Linux clusters. The Deep Computing team will work with IBM’s research groups to connect relevant offerings from throughout IBM’s portfolio. Turek said one of the chief reasons for the creation of Deep Computing is the potential for tapping areas such as cell phone design, medical simulation, Hollywood animation and fraud detection.

Turek, now vice president of the Deep Computing division, dismissed the competitive argument and the same-month timing of the announcements, saying that no other firm can match the scale, scope and breadth of what IBM has to offer.

“One can go into battle with a cap gun and one can go into battle with a bazooka,” Turek told “Each has a gun, but which one is going to win?”

Through all of this, Turek said IBM wanted to steer the perception of HPTC away from the traditional server-only line of thinking, and align its strategy to better fit with the company’s overarching on-demand computing strategy. To be sure, IBM has already announcedsupercomputing on-demand.

IBM’s announced Deep Computing amid its pledge of support for AMD’s Opteron architecture, a launch that has Sun equally excited about its possibilities for HPTC.

Sun’s Kahn offered a frank assessment of the server chip space, vis-á-vis AMD’s Opteron announcement, saying that it further dilutes Itanium’s entry into the server market. He said 32-way architectures have been established, while the 64-way is still up for grabs.

“Right now, there is Itanium, Opteron, and what Intel calls Yamhill, so there are three different sets of instructions [for server vendors],” Khan said. “Intel has the right company but the wrong architecture, AMD has the right architecture but the wrong company,” and it remains to be seen what Yamhill is. Khan said the combinations of Solaris and x86allow Sun to go “from smart card to supercomputer.”

Meanwhile, IBM is developing Blue Gene/L, a supercomputer scheduled that will be used to simulate events such as fires or materials aging, among many other large supercomputer contracts it has picked up in the last two years.

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles