Remember IBM? You know, the company that built a computer that clobbered a couple of all-time Jeopardy champs earlier this year? I know, IBM’s not exactly Apple in terms of its consumer branding, but they do still make a bunch of stuff for mostly business types. Like…okay, I can’t think of anything off the top of my head either, but they do.
Except not “innovative” petaflops-crunching supercomputers, or at least not anymore. IBM and the University of Illinois based National Center for Supercomputing Applications (NCSA) had been partnered to build a petaflops-speed supercomputer, dubbed “Blue Waters,” but they both just backed out of the bargain, citing higher-than-expected expenses and unanticipated complexities.
(LIST: Top 10 Man-vs.-Machine Moments)
One petaflop is analogous to 1,000 trillion floating point operations per second. By comparison, a U.S. Air Force custom-rigged defense computer consisting of 1,760 PlayStation 3 consoles can only crunch 500 trillion floating point operations per second (I know, “only”). Sure, the world’s fastest computer, the Japanese “K” computer, will be capable of 10 peta (or 10 to the 16th power, something like eight quadrillion calculations per second) when it’s completed next year, but then today’s high-end Intel processors can only manage around a comparably trifling 100 gigaflops.
IBM cracked the petaflops barrier back in May 2008 with a computer nicknamed “Roadrunner” designed for the U.S. Department of Energy’s National Nuclear Security Administration. Roadrunner, which can process 1.026 quadrillion calculations per second—consisted of 6,948 dual-core AMD Opteron chips, 12,960 Cell engines (think PS3 processors) and 80 terabytes of memory. So it’s not like IBM’s never built one of these things before.
But times and technology change. Blue Waters, the company’s joint supercomputing endeavor with NCSA, was to be a four-year project (started in 2008) with an estimated price tag of $208 million. According to a joint IBM/NCSA press statement, the University of Illinois and NCSA picked IBM in 2007 as the supercomputer vendor to tackle the Blue Waters project based on “projections of future technology development.” But somewhere along the line, the project went off the tracks.
“The innovative technology that IBM ultimately developed was more complex and required significantly increased financial and technical support by IBM beyond its original expectations,” reads the IBM/NCSA statement. “NCSA and IBM worked closely 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.”
Bummer. And double-bummer for both the IBM and NCSA: IBM has to return all the money it’s received for the project to date, while the NCSA has to return all of the equipment it’s received from IBM.