Intel licensing to AMD
Albert K. wrote:
"I do agree with you 100% but you have to realize that if Intel's clients can not live with intel's business practice then intel is out of business. IBM would never accept single source for the chips and that was it. As far as I recall there is nothing in Constitution that would force Intel on the customers either. :-))"
Agreed. But establishing "second sources" used to be more important than it is today. This is for various reasons, the most important of them being that chip companies often "lost their process" and could not deliver. A friend of mine had some of his key products wiped out because a key supplier couldn't deliver (after having already delivered in the past).
Intel has fewer formal second source agreements than in times past. I may be wrong, but I believe nothing AMD is now producing is _formally_ part of a second-source agreement.
"I found it curious that you take a pride in
"argue like hell before and during a meeting, but once a decision is reached, support that decision.""
Again, what I say and what you say I say are two different things. I did not say I 'take pride" in this, just that is has proven to be effective. More on this below.
"It does not even sound good. Already on a number of occasions Intel suffered because of the lack of flexibility."
It works pretty well. As history obviously shows. Unless you have worked in a large company with many competing interests, which I believe you have _not_ (at least not for a company like Intel), you cannot fully appreciate the "empire building" and backbiting which can and does occur in large, competitive tech companies.
The Intel strategy, which reports indicate spread in part to AMD, worked well.
" Rambus alone is a great example how intel is sticking with decision made no matter what. AMD would not be able to gain that much of a market share and now threaten commercial sector if Intel would abandon rambus long time ago. "
The larger, and much more important and interesting, issue is not that AMD has managed to compete in some markets with Intel but that Intel has so FEW competitors in its core businesses.
I think I wrote about this point in another article recently, so I lack the energy to spend to 20 minutes describing this.
"I am not sure where "copy exactly" was adopted at amd. Amd did have issues at implementing technology but I do not believe that Intel's "copy exactly" was used to solve it. There is too much difference between those companies."
I never said it was used _exactly_.
I remember when a bunch of very talented expats from T.I. arrived at Intel in the late 70s. They came from a culture where each fab manager decided how chips should be made. Well, these talented T.I. ex-pats tried this at Intel. In Oregon and in Albuquerque.
They learned, or were forced to learn, that fab processes are too critical and too sensitive to let each fab decide what to do.
Hence the "copy exactly" methods instituted by CRB after he took over in the mid-80s. (And mostly already in place before that: compared to other companies, Intel was already "McIntel," a term coined by ASG in the mid-70s.)
It is obviously fascist in some ways, from the point of view of a creative engineer, but it clearly works.
AMD's Dresden plant benefitted from this point of view.
--Tim May
Regards |
