ICD 7 Results Thread

[IC Diamond 7]

The calculation seems to make more sense to me.

I am looking at some of the results over different forums. And what popped out in a few instances was at a stock speed you would get a high delta temp. When overclocked the delta would narrow or disappear which is not what happens in real life, the thermal resistance of a material does not change the spread would only get larger with more watts. This jives with this guys description of errors in different programs.

It would be interesting to compare programs temps with a thermocouple embedded in the IHS against absolute temps and across a range of power adjustments

[polobunny]

Originally Posted by IC Diamond 7

The calculation seems to make more sense to me.

I am looking at some of the results over different forums. And what popped out in a few instances was at a stock speed you would get a high delta temp. When overclocked the delta would narrow or disappear which is not what happens in real life, the thermal resistance of a material does not change the spread would only get larger with more watts. This jives with this guys description of errors in different programs.

It would be interesting to compare programs temps with a thermocouple embedded in the IHS against absolute temps and across a range of power adjustments

Yep. I've been wanting to know my E8400 temperatures for a while already. :P

[oldman_gamer]

I will be over sons house tomorrow.....after work,and will get some temps from his Quad core build.I used IC7D on his HSF.I havent had much time to redo my HSF...but will.

[BrokenHalo83]

There's a 10C change from Core Temp to Real Temp. Which one should I believe then?

[One4yu2c]

Originally Posted by BrokenHalo83

There's a 10C change from Core Temp to Real Temp. Which one should I believe then?

It's difficult to tell, and highlights the problem inherent with software temp monitoring. These programs are essentially guessing the processor's un-published TjMax, which is then used to calculate a temperature reading. Problem is, Intel doesn't disclose this information.

Real-Temp's developer attempts to side-step the limitation by using an IR thermometer to better guesstimate a true TjMax, and gives the end user the ability to adjust this measurement, but some with argue this method is flawed. The question is, is CoreTemp more flawed? Very possibly.

[screwballl]

I get the same temps no matter what program I use but I assume that the E6600 B2 revision does not have this problem or they are all reporting the same wrong temp...

[Reloadron]

Originally Posted by One4yu2c

It's difficult to tell, and highlights the problem inherent with software temp monitoring. These programs are essentially guessing the processor's un-published TjMax, which is then used to calculate a temperature reading. Problem is, Intel doesn't disclose this information.

Real-Temp's developer attempts to side-step the limitation by using an IR thermometer to better guesstimate a true TjMax, and gives the end user the ability to adjust this measurement, but some with argue this method is flawed. The question is, is CoreTemp more flawed? Very possibly.

For those with a surplus of time on their hands (about the time it takes to drink a beer) the following is a good read:

Intel® Software Network - IA32_TEMPERATURE_TARGET MSR not documented

Yes it is boring, I never said it was exciting stuff. However, note how the thread evolves till it is locked. Igor as well as several posters to Intel were pretty dedicated to getting an answer. Note the code Igor posted.

Ron

[IC Diamond 7]

Originally Posted by Reloadron

For those with a surplus of time on their hands (about the time it takes to drink a beer) the following is a good read:

Intel® Software Network - IA32_TEMPERATURE_TARGET MSR not documented

Yes it is boring, I never said it was exciting stuff. However, note how the thread evolves till it is locked. Igor as well as several posters to Intel were pretty dedicated to getting an answer. Note the code Igor posted.

Ron

Let me chew on that one awhile. I am thinking this through. thanks for the link

[oldman_gamer]

Ok..here are my sons readings using speedfan after running 3 complete virus scans from three different companies.Best I can do at this time...the readings look great though.
Q6600 (all stock settings on his build)

[Reloadron]

Originally Posted by oldman_gamer

Ok..here are my sons readings using speedfan after running 3 complete virus scans from three different companies.Best I can do at this time...the readings look great though.
Q6600 (all stock settings on his build)

Uh, yes the readings look great, but a little too great. Now I forget but "think" you have a Q6600 in that computer?

Those readings are awfully low (the core temps). My guess is that speed fan is lying to you a little.

Here is something interesting. My laptop is sitting here (Core Duo) running Speed Fan and Core Temp. Interesting that Speed Fan shows me Core 0 and Core 1 as about 35 or 36 Degrees C while Core Temp is showing me about 52 to 54 Deg. C.

Want to try an experiment? Download and run Core Temp and Speed Fan at the same time. I place more stock in Core Temp on a Q6600 processor.

<EDIT> OK on a side note, Kathy got off her computer. She is running a Q6600 and I see about the same thing. Speed fan shows all four cores to be about 15 Degrees C. below what Core Temp shows me for the same cores. That runs with just about what I see on my laptop. Interesting also Speed Fan shows a CPU temp in excess of 100 Deg. C. on Kathy's machine so I can assume it is dead even though it runs just fine. My guess is that Speed Fan can't measure her system with any accuracy. </EDIT>

<Edit Yet Again> Damn, how could I be so stupid! (Don't answer that). Note I see a 15 Deg. C. delta within +/- 1 deg. C. Core Temp sees the correct Tj.Max as 100 Deg. C. for the processors. I will bet that Speed Fan is applying an 85 deg. C Tj.Max? Somewhere in that earlier link Intel does call out which processors use each Tj.Max. Just a guess but that 15 deg. C. is a hell of a coincidence.

Ron