A battle is brewing for the hearts, minds, and wallets of the computer industry that features some familiar companies gearing up, once again, for war. But as we slowly move to 64-bit computing, the expected winner, Intel, has stumbled badly, with its Itanium family selling poorly and underwhelming potential customers. Meanwhile, perennial runner-up AMD has unleashed plans for a 64-bit successor to its Intel x86-compatible chips; the new chip provides all the benefits of a 64-bit code base without any of the compatibility issues the Itanium faces. Determining how the market might respond to an epic battle such as this would usually be difficult, but this time, subtle clues indicate that Intel is about to be handed defeat. And if I'm reading these clues correctly, this battle hinges largely on the opinion of one person—David Cutler.
Most UPDATE readers have probably never heard of Cutler, but he's had a profound effect on anyone who's used or administered a Windows NT-based product. Cutler is the original architect of NT. He's an ex-Digital Equipment Corporation (DEC) veteran who had previously spearheaded the development of VMS but then left the company in late 1988 after DEC cancelled his next OS project, code-named Prism. Developers designed both VMS and NT to overcome the weaknesses of UNIX, which Cutler once described as a "junk OS designed by a committee of Ph.D.s." NT's original design—especially its hardware abstraction design paradigm, multiplatform portability, and kernel—was Cutler's baby, and although he's never sought (or welcomed) any press coverage during his tenure at Microsoft, Cutler is the man most responsible for making NT happen.
Coming from the mainframe world, Cutler's never been a fan of the PC. He considered the Intel x86 line of microprocessors and the OSs that relied on that chip family, to be something of a joke. So in the early days of NT development, circa 1988 to 1990, Cutler focused on a RISC-based chip from MIPS and demanded that his programmers write code that would work on any processor, rather than Intel x86-specific code, which might have been faster but would have been less portable. Cutler's disdain for the x86 continued through the 1990s, as Microsoft ported its NT code-base to other platforms, including DEC's Alpha and IBM's PowerPC.
Of course, NT's cross-platform capabilities didn't seem to matter much by the late 1990s as Microsoft canceled almost every non-x86 port. Cutler's beloved MIPS chip fell first, followed by the PowerPC, and, finally, the Alpha, which had been sold as part of DEC to Compaq. But during this time, Cutler transitioned to a different role at Microsoft, where he continued his work on cross-platform NT. When I asked about him at an NT 5.0 (Windows 2000) Beta 2 technical workshop in August 1998, I learned that he was working on 64-bit NT versions. I assumed at the time that this meant he was working on porting 64-bit NT to the Alpha, but a year later Microsoft killed that port. Naturally, one might gather that Cutler had actually been working on porting NT/Win2K to Intel's Itanium, which was just then starting to bloom. But now I'm not so sure. Cutler has never said anything publicly about Itanium. However, the press-shy Cutler provided a quote when AMD announced plans for its own 64-bit product, which will be called Opteron when the chip begins shipping later this year.
The quote appears in a May AMD press release. "At Microsoft, our vision for 64-bit computing is a highly scalable and affordable platform that is easy to deploy, easy to manage, and easy to develop applications for," said Dave Cutler, Sr. Distinguished Engineer, Microsoft's Windows Team, and one of the world's foremost software architects. "AMD's 8th-generation architecture gives customers great 32-bit performance and 64-bit capabilities on a single system. Together, AMD's 8th-generation processors and Windows should provide customers a flexible platform and a compelling value proposition."
Why is this quote surprising? Knowing Cutler's history, one would logically assume that he'd be more excited by Intel's Itanium, which eschews x86 compatibility for a new design that is supposedly technically superior to any previous chip designs. (Some x86 software still runs on the Itanium, using a Windows on Windows—WoW—like emulation layer.) The Opteron, meanwhile, retains x86 compatibility and adds 64-bit coding extensions that AMD dubs x86-64. From a technical standpoint, x86-64 is sort of a hack, especially when you consider that the x86 has humble technical roots and numerous limitations. But Cutler not only backed the Opteron, he did so publicly—an extremely rare occurrence.
Microsoft can support 64-bit operations in the company's existing 32-bit Windows products, which means Microsoft doesn't have to release a special Opteron-compatible Windows version, as Microsoft does for Itanium. As a result, the company can support Opteron more cheaply and efficiently.
Intel is so nervous about the Opteron that the company has launched its own secretive x86/64-bit hybrid design, code-named Yamhill, that the company will launch if its partners finally give up on Itanium, or the Opteron achieves significant market momentum. Frankly, Intel has a lot to worry about. With Itanium selling far below estimates—some analysts place total chip sales at several thousand units or less—and NT architect David Cutler backing a product that appears to be technologically inferior, Intel's future market position is in serious danger. Does AMD have what it takes to seize the top spot and topple one of the computer industry's most enduring and seemingly unbeatable powerhouses?