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 (82278)	12/7/1999 10:11:00 PM
From: Mani1	Read Replies (2) \| Respond to of 1576422

Bill Re <<The calorimetric method you suggest would be less accurate>> The method I just told you is extremely accurate and fairly fast. I think it is the most accurate way to determine the power dissipation of a chip. Re <<used for heat sink designs to see what it steady states at for different speeds and voltages.>> Yes, exactly. Thermal design is the reason AMD quoted those power numbers in their spec sheet. I by no mean, meant to suggest that it is used for "binning" data. I have no idea how that is done. Mani

To: Bill Jackson who wrote (82278)	12/7/1999 10:13:00 PM
From: Elmer	Read Replies (2) \| Respond to of 1576422

Re: "Mani, You could probably use Ohm's law. After all with more than 20 million transistors they will not all go on at the same time and then off with the clock so you will get an avaeraged effect and I bet if you hooked up a gigahertz scope to the part you would not see anything more than a little grass that represents the power supply surges of those nanosecond transitions that are all averaged out. Since it all ends up as ohmic heating I think you would be within 1% of the truth. The calorimetric method you suggest would be less accurate and would take more time and would be used for heat sink designs to see what it steady states at for different speeds and voltages. A series of experiments in a test jug with a special on chip thermal probe would probably be used for this." Bill this is all nonsense. If you are going to produce 20Million (AMD) to 100Million (Intel) Processors a year how are you going to guarantee the power dissipation to your customers and do it in the same test you are using to guarantee processor operation and performance and you only have a few seconds to do all of it? Answer: you're going to average the current and compare it to the Icc spec published in your data sheet. That's what you guarantee because that's the only way you can measure it in a production environment. We'd all love to hook up scopes but just try it on 100 million parts and see if you meet your quotas. Modern testers average current. EP