Steve, >>>If what I am saying is true, the quicker the processors, the closer
they are at hitting the "brick" wall. Only Intel will have the technical and financial
resources to profitably climb over this wall, and make it a one horse race once
again.<<<
I agree with what you say, i.e., the cost of entry and continued participation in the computer hardware market is huge, and not many can afford it. I mean participate and really make a difference and dominate, not having a token x86 like IDT's Winchip. When you add up the cost of architecture, design, simulation, verification, diagnostics, software, process engineering, manufacturing and everything else, the cost is easily in the billions for a significant new chip like Merced. Who else can afford to do it and do it right, i.e., come out with a new chip or system? Intel, IBM, Sun, a Japan-US alliance or two, and AMD, I'll say, valiantly trying. Even Hitachi, a $38 Billion company threw in the towel and signed with IBM to use big Blue's S390 chipset. Hitachi used to do all their own CPU logic and chip design. Not any more.
As far as possibly not being able use all the megahertz Intel can crank out in PCs (hit the wall, which I don't believe) there is absolutely no worry about that for the medium-high end workstation and server chips Intel is also majoring in now. Xeon, Merced etc., will be appreciated for every last MHz and cache byte Intel can muster. Speed kills (the competition) up where these chips play.
Tony |
