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: John F. Dowd who wrote (122993)	8/23/2000 9:19:29 PM
From: Dan3	Read Replies (1) \| Respond to of 1578494

Re: 80 degrees C = 176 degrees F. Is this really a problem? Where has this been documented? Sounds like FUD to me. JFD Hi John, The temperature inside the case is usually 10 to 20 degrees above ambient. And most OEMs consider 80C OK. But the fastest current parts have a max at 62C - a little less than 144 Fahrenheit. Think of one of those tiny 40 watt light halogen bulbs inside of a box. Mount it on a thermally insulating fiberglass board so all the heat must be removed from one side. Now make it an absolute requirement that the bulb never ever get warmer than 144 degrees. That's the situation with Intel's 1GHZ and higher processors (the original 1GHZ P3s had a 60C max!) Even keeping an Athlon below 90C (194F) requires a decent fan. But there's a lot more margin with the Athlon - everything doesn't have to be perfect. Here's a copy of an earlier post of mine: Dan ============================================ The thermal requirements for Thunderbird haven't changed. The spec still tolerates a max 90 degree centigrade CPU temperature at all speeds - no special requirement for the higher speed processors. Plenty of headroom - a nice, robust chip. amd.com Meanwhile, the latest stepping of the PIII 700-866MHZ chips will tolerate 80 degrees, but the 933 requires cooling (to 5 degress C less than a standard P3), the 1.0 GHZ requires special cooling (to 10 degrees C below a standard P3) and 1.13 GHZ basically needs a kryotech case (requiring cooling to 18 degrees C below a standard p3). developer.intel.com It appears that, despite all the publicity, Intel's C0 stepping has failed to bring coppermine binsplits anywhere near to AMD's speed grades. Interestingly, the spectre of Duron seems to have collapsed PIII pricing for parts running 733 and lower. PIII, Celeron, Duron, and K6 systems all seem to sell for about the same price at a given MHZ where each is available. Message 14220656

To: John F. Dowd who wrote (122993)	8/23/2000 9:25:57 PM
From: combjelly	Respond to of 1578494

80 degrees C is not FUD. The fact that Intel had to drop the temperature requirement is a sign (like raising the voltage) that the chip is being overclocked beyond the design specifications. If the AMD chips can take a higher temperature (even if it requires a higher voltage) that means they have more headroom.

To: John F. Dowd who wrote (122993)	8/24/2000 2:37:22 PM
From: Scumbria	Read Replies (2) \| Respond to of 1578494

John, what is so different about the AMD chip that it can take the heat? Semiconductors generally have a relationship between temperature and speed, where the higher the temperature, the lower the speed at which the part will function correctly. Normally, parts are tested at 80-90C because that is the upper limit of temperatures found in normal operating environments. All PIII parts which run below 933MHz are rated up to at least 80C. Intel is unable to manufacture PIII's that run at 1GHz and 80C, so they lowered the operating temperature to 60C. This is a very questionable practice, to say the least. If AMD chose to lower their standards to 60C, they could be shipping parts at 1.4GHz right now. I hope the company has the integrity not to do that, however. Scumbria

To: John F. Dowd who wrote (122993)	8/25/2000 1:32:52 AM
From: jcholewa	Respond to of 1578494

> And what is so different about the AMD chip that it can take the heat? Sounds like FUD to me One factor is that it has a larger die area, I'm told. More area over which to dissipate its heat means less heat per unit area (of course, it still bunches up in areas, prolly the center, but I suppose the basic axiom holds at least a little). There is likely some point to what the people here are saying. To me, it's really just a matter of companies taking their parts out of spec. I've given up on complaining about voltage hikes, because they've become totally commonplace for both Intel and AMD, but this whole junction temperature thingy sounds at least slightly fishy. Of course, I'm probably just riled up since Intel just did three 'K6-266' moves in a row (eg, announcing a product, then leaving half a year before the product actually becomes reasonably available) this year. That sort of particular action unnerves me. -JC