CoreTemp 0.96.1??? New Temp reading for M0 Core2Duo?

[GundamF91]

I remember when the CoreTemp 0.96 came out, it raised the M0 TjMax reading to 100c. Supposedly that was needed because Intel raised the Tjunction Max.

Now CoreTemp people said it's been fixed once again, and the Tj Max for M0 is back down to 85c again. So what's up with that?

There was lots of talk of adding 15c to the CoreTemp and Speedfan readings, but if the latest fix is the correct reading, then it would appear the original temp was correct. There's no need to add 15c to M0 and G0 processors?

 

[Mondoman]

CoreTemp 0.96.1 is reading my e2160 M0's temps correctly (in that they are compatible with the MB's external CPU temp sensor's readings).

 

[GundamF91]

What's your temp Mondoman? My E4500 runs 23c idle, that's based on both motherboard BIOs as well as CoreTemp 0.96.1 reading. But CoreTemp 0.96 used to report it about 15c higher, at 38c. Funny thing is that my motherboard temp is around 35c, so my CPU is always running below my mobo temp. Now I'm not sure what's the real temp.

 

[Mondoman]

I'm in the middle of running some long Orthos runs at different FSB speeds, but at idle in XP CoreTemp reports low 20s C, and low-to-mid 50s C during Orthos, depending on FSB speed. The MB CPU temp sensor reports temps roughly 2-5C lower, with much less fluctuation. The MB "System" temp is reported at 30C during Orthos. This is with the stock Intel cooler running in its temp-sensitive mode.

 

[Pancake106]

im in the same situation! i have a m0 e2140 at 3.2ghz and 1.38v with a hyper tx2 cooler. my temps are a tad higher than you guys at idle and load. im confused on the whole readings but coretemp 96.1 sure makes me feel better even if it is infact off lol. can we compare m0 temps to older chips to figure out which is closer and possibly more accurate???

 

[soccerballtux]

Originally posted by: Pancake106
im in the same situation! i have a m0 e2140 at 3.2ghz and 1.38v with a hyper tx2 cooler. my temps are a tad higher than you guys at idle and load. im confused on the whole readings but coretemp 96.1 sure makes me feel better even if it is infact off lol. can we compare m0 temps to older chips to figure out which is closer and possibly more accurate???

Get Speedfan 4.34 beta; whatever temps those are is correct.

After CT 0.96.1 these two readings match; my M0's readings were 15c too high in 0.96.

The previous readings were wrong because I was pushing my core to 74C. If the real temp were 89C the motherboard would be screaming and would shut off the proc the second it hit 70.

 

[VirtualLarry]

Originally posted by: soccerballtux

Originally posted by: Pancake106
im in the same situation! i have a m0 e2140 at 3.2ghz and 1.38v with a hyper tx2 cooler. my temps are a tad higher than you guys at idle and load. im confused on the whole readings but coretemp 96.1 sure makes me feel better even if it is infact off lol. can we compare m0 temps to older chips to figure out which is closer and possibly more accurate???

Get Speedfan 4.34 beta; whatever temps those are is correct.

After CT 0.96.1 these two readings match; my M0's readings were 15c too high in 0.96.

The previous readings were wrong because I was pushing my core to 74C. If the real temp were 89C the motherboard would be screaming and would shut off the proc the second it hit 70.

How many times do I have to explain this to you. The thermal shutdown on the mobo depends on the measured Tcase temps. CoreTemp measures Tcore temps. They are not the same. Generally, TCore is 15C higher than Tcase. So with your core temps at 89C, that would mean your Tcase temps would be around 74C. If your mobo shutdown temp were 80C, then it wouldn't reach that.

See my other thread for the reason that CoreTemp 0.96.1 is INCORRECT about core temps.

If the TJmax of an M0-step chip really were 85C, then those chips should throttle at that temp.

Yet, with CoreTemp 0.96, I measured 85C core temps on my M0-step E2140. It was NOT throttling. Therefore, the TJmax must indeed be 100C for these chips.

Thus CoreTemp 0.96.1 is BROKEN.

 

[VirtualLarry]

Originally posted by: Mondoman
CoreTemp 0.96.1 is reading my e2160 M0's temps correctly (in that they are compatible with the MB's external CPU temp sensor's readings).

Again, this is PROOF that CoreTemp is WRONG.

The mobo's temp sensors are reading Tcase (approximately). CoreTemp measures Tcore. Tcore should be 15C HOTTER than Tcase.

If the readings are roughly the same, then that means that the Tcore reading is 15C TOO LOW.

CoreTemp 0.96.1 is BUSTED.

 

[VirtualLarry]

Originally posted by: GundamF91
What's your temp Mondoman? My E4500 runs 23c idle, that's based on both motherboard BIOs as well as CoreTemp 0.96.1 reading. But CoreTemp 0.96 used to report it about 15c higher, at 38c. Funny thing is that my motherboard temp is around 35c, so my CPU is always running below my mobo temp. Now I'm not sure what's the real temp.

Another example of CoreTemp 0.96.1 being off. It's impossible for Tcore to be less than Tcase.

 

[Mondoman]

Originally posted by: VirtualLarry
...
Another example of CoreTemp 0.96.1 being off. It's impossible for Tcore to be less than Tcase.

He wasn't saying that; he said his MB "system" temp sensor reported a temp higher than either the MB "CPU" temp sensor or CoreTemp, which is perfectly reasonable, as the CPU is cooled by an attached active CPU cooler, while the "system" temp sensor is typically not actively cooled and may not have much airflow at all.

 

[Mondoman]

Originally posted by: VirtualLarry
...
If the readings are roughly the same, then that means that the Tcore reading is 15C TOO LOW.

CoreTemp 0.96.1 is BUSTED.

To clarify, when I wrote "consistent with", I did not mean "the same". Rather, my MB sensor readings are roughly 3-5C below the fluctuating CoreTemp 0.96.1 values at steady state, and about 10C or so below CoreTemp during the temp upramp phase (e.g. just after starting Orthos).

 

[Mondoman]

Originally posted by: VirtualLarry
...
The mobo's temp sensors are reading Tcase (approximately). CoreTemp measures Tcore. Tcore should be 15C HOTTER than Tcase....

I agree that Tcore should be hotter than Tcase. However, it's not clear to me where the 15C value comes from. Naively, I would expect the difference to vary with the CPU temp, and likely also among the different core designs. If you have a link/ref with info on what the difference should be, I'd welcome it.

 

[VirtualLarry]

Computronix's sticky thread over at the tomshardware forums documents the 15C delta between Tcase and Tcore. He posted that thread here as well, but after it didn't get a sticky (after only a few days), he erased his post and went off somewhere to sulk.

I've personally determined that value to be empirically correct as well.

 

[Old Hippie]

Any temp probe or program that reports my chip idling in the low to mid 20s on air, is probably incorrect. I've always been @ 32-33C on any chip, and I don't skimp on case air flow. 23C is @ my ambient room temps. (73-74F), and with any power going thru the chip, it's gonna heat up a little. I'm not saying it's impossible to hold 23C, just not probable...at least for me.

So Larry, why did they "fix" Coretemp 96.1?

 

[VirtualLarry]

Originally posted by: Old Hippie
So Larry, why did they "fix" Coretemp 96.1?

Too many whiners? Unbelieving that a piece of silicon could actually get that hot and still operate? (GPUs do it all the time, btw.)

 

[soccerballtux]

Well, no matter what, I'm only exceeding Intel's Tcase temp recommendations by 1degree C worst case when loading small Orthos; so my proc will be fine.

 

[Binky]

CPU temps reported on various forums have been completely useless since the Core2 came out. I find it hilarious (and annoying) that people actually think that their overclocked 3.xghz cpu on 1.4v+ is idling at 1C over room temp. Seriously, smack yourself in the head if you think this is possible.

 

[Mondoman]

VL - thanks for the pointer to that Temp Guide. I couldn't find any ref/description in there of how they were determined, so I've posted a question there to (hopefully) get the details.

 

[Old Hippie]

I find it hilarious (and annoying) that people actually think that their overclocked 3.xghz cpu on 1.4v+ is idling at 1C over room temp. Seriously, smack yourself in the head if you think this is possible.

Exactly! Some people seem to lack a little common sense.