Pigboy - Re: "Not in this market, where product life cycles are nine months.'' >> Is Ron smoking something? Aside from quality, I would assume size, volume (p r ice) matters just a little bit. Isn't AMD realizing that now? :-)"
There is truth to both sides of this,
Graphics chips companies have historically been small, nimble and agressive - AND Fabless!.
Intel, as a large company, has near-infinite layers of management and all these layers need to interact, to "buy-in" to decisions, and all that other corporate garbage.
The graphics group will have to operate in an Intel-Specific mode - addressing key, strategic, long range objectives, or ELSE operate autonomously from the corporate legions - in order to make fast, quick decisions and implement them.
Intel's fab capacity is "negotiated" even among existing groups - the chip set designers have to fight for wafer start allocations on the newest processes even BEFORE they can begin circuit designs. Else, they are relegated to use the older, depreciated fab lines and processes.
A good example is the i740 Chip. For a number of reasons, this chip was designed and implemented in a 0.6 micron process! This process is 5 1/2 years old!
The chip was slow - but not that slow - and sucked enornormous power! Many references to this have been made in the media.
However, after verifying most of the chip's functionality, it was relaid out on the 0.35 micron process about 6 months ago and that is the part that will be rolled out in the next few weeks or so.
This i740 chip isn't Intel's first graphics chip. They introduced the 82786 graphics processor (originally referred to as the "P6") in the mid 1980's. It was steamrollered by TI's graphics chip at the time.
Paul |
