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: Aaron Cooperband who wrote (51134)	2/26/1999 7:07:00 PM
From: Elmer	Read Replies (1) \| Respond to of 1571808

Re: "So, for example, on a 25 micron process its OK to put any voltage through the circuits? I have to believe there is some limit thats prudent for a given line size. This can be the constant that allows a comparison." I believe Yousef addressed this point. Elevated voltage accross the gate risks 2 things. #1 Actual breakdown of the gate oxide, unlikely here but more likely that if it were run at 2.2v. #2. A upward shift in vt's due to hot carrier injection which will degrade performance over time making in effect, a 450Mhz device into a 433Mhz device or some other such degradation after weeks or months of operation. What Yousef might have add is that this can be compensated for at test where for example a 450Mhz part must run at 475Mhz to be acceptable, considering the HCI affects that will take place with time. This is considered an act of desperation in the industry. But if you have no other choice you do what you have to do to stay alive. EP

To: Aaron Cooperband who wrote (51134)	2/26/1999 7:16:00 PM
From: TGPTNDR	Read Replies (2) \| Respond to of 1571808

Aaron You got into this mire by posting <Re: "It seems to me that if AMD can't yield significant quantities at 400Mhz but Intel can at 500Mhz, doesn't this put them more than 2 speed bins away?"> It looks like AMD can yield significant quantities at 400Mhz, and 450MHz, and that their 400 MHz processor is faster in most important respects than the INTC 500 MHz processor. Doesn't that leave INTC at least one speed grade behind in process technology? How would you limit voltage? The manufacturers rate by defects over lifetime. How would you rate efficiency in a car if the fuel pressure at the injectors varied by - say 150 PSI. It is a meaningless question. The design limits the voltage and the fuel pressure. Youseless well knows that in the December mask fix the voltage path increase was the chosen method, and that the reliability will be there. tgptndr

To: Aaron Cooperband who wrote (51134)	2/26/1999 7:36:00 PM
From: kash johal	Respond to of 1571808

Aaron, Re: Comparing process Aaron you simply cannot compare the 2 processess just based on supply voltage. For example Intel has several variations of its 0.25 micron process as does AMD. Clock frequency is one comparison although still rather meaningless. For example the 400Mhz K6-2 is faster than the PIII on certain tasks and slower on others. That's why there are all these benchmarks such as at Tom Pabst's site and Anands etc. attempting to do a performance comparison by using similar hardware and similar software programs. In general the KIII looks very competitive to a PIII in performance for common business applications. In certain gaming apps and some forthcoming SSE apps I am sure the PIII will show a clear lead. The FUD regarding the voltage issue is just that from some of the Intel faithfull. Regarding how many speed grades AMD is behind Intel the answer is two speed grades. The 450's from AMD were just announced and we will know in a few weeks if they can really manufacture these in volume. At that point AMD will be one clock speed behind. AMD may of course catch up and possibly overtake Intel in Q2 by release of the K7. Although with AMD's poor history of execution this is highly unlikely. Regards, Kash