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: THE WATSONYOUTH who wrote (91680)	2/5/2000 12:04:00 PM
From: Elmer	Respond to of 1576350

Re: "Where do you get 12% HE guardband????. Either you can't read or you misunderstand what I have said. " I got it from you: "More indication that Intel may be using a guard band as large as 12%" The context of previous discussions was Hot Electron guardbanding so it appeared as though that was what you meant. If you didn't mean "HE" but instead "total" GB it was not very obvious from your post, sorry. In any event the guardband added by the test environment is small compared to the effect of the distance between the bins and the HE GB. Your second analysis was good and certainly makes much more sense than the first. Thanks for the followup. Re: "what do you think a reasonable tester guard band should be and why???" Generically, the tester GB would be based on the tester's accuracy. Perhaps 100ps for the one Intel is using. Maybe a little better. Using a HE GB of 5%, that would make the basic test rate ~140MHz or 7.14 ns cycle time. A 100ps tester GB would amount to only about 1.4%, maybe less if the actual accuracy is 80ps. There is debate as to using 2X the tester accuracy but I don't think Intel is doing that. Do they need to GB for untested speed paths? I don't think so because I think they know exactly where the limiters are based on painstaking speedpath analysis on real silicon. I would expect AMD is in the same position. Cheers EP