Conclusion
A good argument could be made that despite a promising start with the original POWER series, IBM has failed to execute up to its potential in the high end RISC processor market. As a result it was as vulnerable to encroachment by Intel’s future IA-64 product line as the other independent MPU players (like Compaq and Sun) despite its awesome semiconductor technological development and manufacturing capabilities. The POWER4 is a rare technological tour de force that simultaneously pushes forward the state of the art on many different levels: superscalar processor design, chip level multiprocessing (CMP), interprocessor communications scalability, and innovation in large system physical packaging. IBM must be doubly applauded in an age when huge schedule overruns are almost the norm for ambitious high end microprocessor projects, for it appears they have encountered no serious delay or setback in the POWER4 development program. First pass silicon has been running since January and has proven sufficiently functional to serve as the computational platform for continued engineering development in the product line.
Although POWER4 was likely primarily intended to take market share away from Sun Microsystems and defend the server high ground from Intel’s x86 and IA-64 barbarian hordes, it appears that the Compaq Alpha family may be the most vulnerable. The Alpha has neither the deep pockets that support IA-64 nor a software installed base as broad and valuable as SPARC. The Alpha’s primary attraction is its perception of unparalleled technical excellence and performance. The POWER4 will eclipse the EV68 and perhaps even the EV7 in many aspects of both uniprocessor and system level performance. It is important for Compaq to bring the EV7 and EV8 to market as quickly as possible without any of the missteps and lost opportunities that distinguished the roll out of third generation products from the first two.
Sun Microsystems has adroitly exploited its software base and sterling image in large corporate IS departments to overcome its customary position of trailing the high end RISC pack in processor performance and achieve great commercial success. Although it will likely have refreshed its large-scale system products with the UltraSPARC-III by the time the first POWER4 systems ship, Sun’s offerings will probably still be clearly outclassed. Worst yet for Sun, IBM is one competitor whose credibility in the corporate corner office exceeds even its own. Sun will need a healthy dose of Andy Grove’s self-attributed paranoia in the coming years to help keep close watch on its competitor from Poughkeepsie, as well as those inspired from Santa Clara, as it tries to hold onto its market share. Conversely, Intel faces a tremendous uphill battle to establish IA-64 against entrenched competitors with no place left to fall back to. POWER4 complicates that task, but is unlikely to change Intel’s grim determination to push forward at any cost.
Footnotes
Be the first to discuss this article!