Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Bill Jackson who wrote (63253)	6/24/1999 9:33:00 PM
From: Process Boy	Read Replies (1) \| Respond to of 1572185

Bill - <PB, Would they do that optical shrink just for speed increases?, with all that dead area around the chip?? I would think they would do an optical shrink and then close up the steps to minimize that dead area and thus get better yields? Otherwise why do it?, no faster, no better yield, or is it a transitional step towards closing up the steps> I don't believe the .18 K7 is what I call an optical shrink. I don't believe the .18 K7 was referred to as an optical shrink. I don't believe it would make much sense to do an optical (dumb) shrink on the K7. Maybe Scumbria can shed some light on you questions. PB

To: Bill Jackson who wrote (63253)	6/24/1999 9:48:00 PM
From: kash johal	Read Replies (1) \| Respond to of 1572185

Bill, What PB has described is an interim step. When you introduce a new standard part you just design it and fab and then fix everything and just ship the hell out of it. In CPU's there are continual mask tweaks to get every scrap of performance out of the chip. In addition folks tweak the transistors to see what additional performance can be tweaked. An "easy" first step is just to shrink the transistors via CAD to new transistor specs leaving everything else identical. This reduces a whole bunch of variables and lets you evaluate any transistor performance/yield issues. Once happy with the transistors you can then fix those and relayout the chip with tighter metal rules and get the much smaller die size ut perhaps with not much speed improvement. Once that works you start up on the transistors again and optomize them again for increased speed. In the past AMD has not had the luxury of a proven production process so they have had to throe a new design and a new process together. Clearly with more vaiables a tougher task to master rapidly. This time AMD has a proven 0.25 process with very fast transistors already. So the K7 has been easier to ramp up (so far). Now they are rampig that and in parallell have down the relayout to the 0.18 micron chips. If they do a good job the 0.18 micron parts will not only yield like a bat out of hell but be faster too. On the other hand they could have a surprise like Intel did with its 0.18 parts. So expect a lot of transitional steps and not a giant leap on the K7 ramp. Regards, kash