How does that make those percentage figures bogus?<<<<<<<<<<What is bogus, is the assumption that this plant, or any other plant in the world is producing at 100% of its production capacity, day in and day out, every day of the quarter, without any allowance for down time at even one machine. That is what is ridiculous, Most plants when running full out run at 80% to 90% of capacity, in the good times.
--I spoke to my Industrial Engineer about this, since it is the I.E.'s that generate all those run rate, wafer capacity, blah blah blah models. He said that yes, it's true that there is equipment down time. But that is ALREADY factored into the number of tools you need to buy to support the capacity. If a tool historically has been down 10% of the time, and you calculate a need for 10 tools (perfect situation) to run to capacity, then instead, you would need to buy 11 tools (10 tools x 10% down = 1 extra tool needed). So capacity INCLUDES things like machine down, wafer rework, engineering experiments, etc.
--As for the rest of it, since neither intel NOR AMD releases specific Yield numbers, it's always a guess. But I notice that I don't see you jumping down the throats of any of the people on this thread for their guesses. Of course, their guesses usually support AMD, don't they?
it doesn't say anything about yield on that process at all.<<<<<<<<<True, but do you really think that semiconductor international would have singled out this fab for recognition, if its yields were bad, no matter what the process was.
--Since we have already agreed that AMD's yield numbers are not published to Semiconductor International, Their recognition has NOTHING to do with Yield, despite your now 2nd attempt to make it seem so. Who's being "bogus" now?
In other words, the proof is in the pudding. Amd is hardly the only manufacturer to use the copper process. In fact,AFAIK, IBM built a 100% copper fab also, and spearheaded the copper process.
--Nice try, but IBM did not build a 100% Copper Fab. They did their Cu development in the current Development Fab. Unless you have a link to prove otherwise? And once again, I guess I need to direct your attention to the fact that the award was for "The First", It doesn't say anything about the quality of the process. I doubt that SIA would have awarded "The First" to AMD, if IBM were first.
Why wasn't Intel's notched gates recognized last year, if all Semi, Int. was looking for was a new process. To be the best fab in the world, you have to be good at everything; not just one aspect. And certainly, yields are the most important aspect.
--So, I think I'll give ya a little break here, since you don't seem to understand semiconductor manufaturing. "Notched Gates" were NOT a "Process", it was an improvement to a process that was already in production. Notched gates was a change that happened AFTER P858 was proliferated to the HVM sites. And by the way...... That award (to remind you again), was not for "The Best Fab In The World". That's 3 times now...
Why is it when Elmer can't think of a good reason for something, that automatically means something is desperately wrong at AMD. There are other reasons.
--Ho Ho Ho, sometimes answers are so easy, I almost don't want to take the shot. Sorry, can't resist. I would guess that the reason that Elmer thinks that "something is desperately wrong at AMD", is probably for the same reasons that most (all) AMD fans (including you), automatically think the same thing about intel. So if you want an answer to your question...... ask yourself.
Semi |
