Another vendor being sought for Blue Waters after IBM leaves


Another vendor being sought for Blue Waters after IBM leaves

URBANA — The University of Illinois industry-leading water-cooled supercomputer hit a setback Monday: IBM no longer finds it cost-effective to work with Blue Waters.

A prime supplier, IBM has ended its relationship with the project effective Aug. 6.

IBM spokeswoman Joanna Brewer said the project was getting too expensive.

"The Blue Waters contract was awarded to IBM in 2007 based on preliminary and conceptual design points. Greatly increased costs resulting from the final design, and ongoing changes required during the implementation phase led to the decision by IBM to no longer provide the Blue Waters supercomputer," she said Monday.

She said IBM's Power775 servers are ready and will be shipped to other buyers later this month.

John Melchi, NCSA's senior associate director for administration, said a new major vendor will be found, but it's too early to comment on whom, since the deal fell through only days ago.

Melchi said NCSA is "guardedly optimistic" that the project will continue with no further troubles.

He also said that in discussions with the funding agency, the National Science Foundation, the National Center for Supercomputing Applications has said it will meet its fall 2012 deadline and the supercomputer will reach a sustainable speed of 1 petaflop — one quadrillion floating point operations per second. The NCSA explains that as, "if you could multiply two 14-digit numbers every second, it would take about 31 million years to complete the 1 quadrillion calculations Blue Waters will complete every second."

According to a joint release by the company and the NCSA, IBM will return money received to date. The NCSA will return equipment delivered by IBM according to the terms of the contract.

IBM had to give back $30 million to the NCSA, according to the high-tech blog at

But the NCSA said it is still going forward with the project, though possibly with a different name, since Blue Waters is a reference to IBM's nickname "Big Blue."

Melchi said no decision will be made on renaming the project at this time.

Lisa-Joy Zgorski, a spokeswoman for the NSF, said the foundation would be reviewing changes.

"Right now it's the status quo. We anticipate that Illinois will propose alternatives, and we expect that any proposed revisions will be within the general scope of the project funded," she said.

She said the NSF expected to be reviewing a new plan by mid-September.

The Blue Waters project is intended to produce a sustained performance of 1 petaflop on a range of science and engineering applications such as predicting the behavior of complex biological systems, understanding how the cosmos evolved after the Big Bang, designing new materials at the atomic level and predicting the behavior of hurricanes and tornadoes, the NCSA has said.

Trish Barker, a spokeswoman for the NCSA, said "IBM has terminated its contract to provide the hardware, but IBM will still be building its (next-generation) Power7 hardware."

"We anticipate quickly trying to replan the project with an alternate type of hardware, on the same timeline of 2012 as if we had been staying with IBM."

"It is certainly our hope that it will not cause further delay," Barker said.

According to a statement, "NCSA is confident that its goal of building a sustained-petascale supercomputer remains achievable in a timely manner. NCSA is coordinating with the National Science Foundation to ensure project continuity and that the goals of the project are achieved."

The UI and the NCSA selected IBM in 2007 as the supercomputer vendor for the Blue Waters project based on projections of future technology development.

"The innovative technology that IBM ultimately developed was more complex and required significantly increased financial and technical support by IBM beyond its original expectations," the NCSA and IBM said in a statement.

The NCSA said it worked closely with IBM on various proposals to retain IBM's participation in the project but could not come to a mutually agreed-on plan concerning the path forward.


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
birdfarmer wrote on August 09, 2011 at 10:08 am

OK, so the NCSA super-geeks have set back this country's lead in supercomputers by four years since the 2007 award for the Blue Water's computer was to be on-line by "summer of 2011", letting the Japanese and Chinese be #1 and 2 in current supercomputer fastest performance. They'll probably all get the maximum "performance-based" raises this year (watch the Salary Guide for details).
Because the State of Illinois has about $100,000,000 invested in the Blue Waters building and infrastructure, we state residents need to come up with a new plan for the building and the $200M of NSF money quickly. How about:
Blue Danube: A big dance hall with LOTS of disco balls;
Muddy Waters: a stormwater detention building, saving valuable below-ground real estate;
Green Waters: buy $200,000,000 worth of Chevy Volts and loan them out to the community, charging the batteries all at once at night;
Rainbow Waters: a new GLBT center;
Brown Waters: rent the building out to UPS as a distribution center????

Alexander wrote on August 09, 2011 at 11:08 am

I don't think they "allowed" anyone to be #1 and #2; that's scientific competition. Maybe the reason why China and Japan are doing so well is that their citizenry support scientific and engineering innovation instead of incessantly complaining at the first opportunity.

birdfarmer wrote on August 09, 2011 at 12:08 pm

The NCSA was given $200,000,000 from the NSF, about $100,000,000 in State support and four years to contract for a computer to be built and become functional. That seems like sufficient financial allowance and certainly not the first opportunity to complain for a competent organization to retake the lead in "scientific and engineering innovation". The problem is not that there is no support for such projects, it is that there is no accountability for complete failure on the part of those managing these large-scale projects.

Alexander wrote on August 09, 2011 at 12:08 pm

(1) The final project is not dead yet; just the partnership (as it was) with IBM. It is not yet 4.5 years.
(2) That there is support from the NSF does not immediately contradict the presence of complaining
(e.g., even before the project timeline is complete).
(3) IBM should be ashamed of itself, IMHO. However, they know they can do things like this since they are the industry leader in computer consulting.

jhhintexas wrote on August 16, 2011 at 6:08 pm

Name one successful company which continues to support a program at a loss. Just ONE. IBM nearly bought the farm in efforts to continue OS2.

They are directly responsible for salaries and benefits of hundreds of thousands of employees and indirectly for even more through supporting business, not to mention the retirement programs and millions of investors who depend on their *revenue* as part of their portfolio [read - retirement]

When projects exceed costs, in spite of efforts to reduce them - they get scrapped. Let's say you agree to roof my home for a negotiated price, and 1/2 way into the job, structural issues beyond your control starting impacting your work, causing you to bring more workers, replace material and repair some of the structure to support the roof. Once you fix that, I expect your crew to be available for questions 24x7, and they charge you overtime. do you continue to work on my roof for the negotiated price?

Alexander wrote on August 16, 2011 at 8:08 pm

"Name one successful company which continues to support a program at a loss. Just ONE. IBM nearly bought the farm in efforts to continue OS2."

I don't need to name ONE. The challenge is irrelevant.

Correct me if I'm wrong -- did they have a contract to complete the job at a fixed price or not?
If I agree to roof you home and find out that I should have charged more half way then unless there's some complex language in the contract then YES I need to finish it. Otherwise that's breach of contract.

I stand by experience: they know they can get away with this shabby consulting work because they're IBM. If they do this to a company, the manager of that company can just say "well, I hired IBM, how was I to know?" That's the advantage of their marketing position.

animal lover wrote on August 09, 2011 at 11:08 am

@Birdfarmer: Go to and give them a piece of your mind.

jhhintexas wrote on August 16, 2011 at 6:08 pm

Are you suggesting that people gripe at IBM for ending a program which will continue to lose money? Seriously?