Originally posted by: lopri
Originally posted by: ArchAngel777
Originally posted by: Binky
With all the strange temps shown on these new cores, I wouldn't believe what you see...
Yep, I even question the G0's revision of the Q6600 temperatures. Some of them idle at just above reported room temperature, which is ridiculously improbable. No one really investigates these issues though. Be nice if a hardware site started using their own micro diodes in these things to see if there is truly a heat difference, or if the revision simply changes what is reported.
I have to say that my attitude is going towards the same way. I've got a Q6600 G0 and running it on 975X, clocked @3.4GHz. Load temp rarely goes past 54~56C with all 4 cores loaded. Well, it's certainly nicer than seeing 70C+, but I don't know if I should believe this temp or not.
It is kind of like religion. Not many people are really interested in researching what they believe. Instead, the take face value and sling it around.
I don't know a lot about microprocessor design, but I do think I have an understanding of heat dissipation and I will attempt to explain why I believe the temperatures cannot really be trusted.
During idle a CPU uses much less power, and therefore generates less heat. During this process only a small fan on the heatsink would be required to dissipate the heat for optimal temperature. Turning the fan to high will not yield better results, because the heat is already being dissipated optimally. However, the minute you turn up the heat by putting the CPU under load, the need to not only asborb the heat goes up, but the need to dissipate it faster is also required, otherwise the chip will become perpetually hotter and will fry itself.
So, let me give two scenarios - now they are made up - be aware, but they should be ballpark figured.
Q6600 B3 Idle = 20 Watts Heat Output
Q6000 G0 IDle = 16 Watts Heat Output
Now, in this scenario, should one expect the the G0 revision to idle cooler? No, not neccessarly. The heat output of both of those is quite low at idle and while the G0 does in fact output less heat, it shouldn't cause temperatures descrepencies that you see here on this forum. If the heat is being dissipated properly, both should idle around the same temperature. Now, one might require the fan to be running at a higher speed (B3) but both have the ability to idle at the same temperature assuming that they are tested in the same environment. The reason this is the case is because when cooling with air, you cannot cool something to a lower temperature than the air you are blowing on it. The best case scenario is being able to cool something to ambient temperature.
But this isn't reasonable for a CPU as you cannot have perfect, metal to metal contact with the Heat Sink and IHS. Even if they were perfect, you then have to have the IHS and the contact with the Cores to contend with. Those are not a perfect fit either and without that perfect fit, you cannot transfer all the heat fom the Cores ---> IHS ---> Heat Sink. So some heat is lost (Not really lost, but just isn't transfered to the Heat Sink) and thus, a running CPU cannot attain ambient temperatures.
So, what really is the deal with the G0's or even the Penryn's for that matter? What possible reason could they be reporting such low idle temperatures?
1) The Diodes in the CPU are not accurate.
2) Some revisions may have poor contact with Core ---> IHS.
3) Some people have poort contact with IHS ---> Heat Sink.
4) Some people just lie.
I personally put more stock into #1 and #2 overall. I think #3 isn't an issue for the vast majority of techs out there and #4 is always an option because people always lie.
Keep in mind for the sake of discussion I am not claiming that a G0 should not run cooler (especially under load). My main point of contention is that they run so much cooler, even under idle (where the delta should be almost non-existant between the B3 and G0) that it seems to be improbable based on their heat output alone. I don't think these things are accurate and I don't think the G0 runs much cooler than a B3. Of course, I can't prove it without testing both and using an independant micro diode. Something I have no plans to do. But I think I have more than enough reason to cast doubt on the G0/Penryn temperature claims.
thoughts?