e2180 oc coretemp vs hardware monitor

[igloo15]

OK so I am ocing my e2180 right now I have a 1.36volt with about an 3ghz overclock (333x9)

Coretemp reads my temperature as 47c while CPUID Hardware Monitor reads it at 62c while running prime95.

Idle the temps are 22c and 37c respectively.

Which should I believe because if its 47c I think i have alittle more headroom to oc more but if I believe Hardware Monitor than I should stop now.

Which temp reading should I believe?

 

[hnzw rui]

Try another utility to measure temps. Speedfan, perhaps?

 

[igloo15]

just tried speed fan its temps appear to match coretemp

Its core0 and core1 temps match coretemp so I am going to guess that coretemp has the correct temperature

 

[QuixoticOne]

Believe both. I think maybe it is reporting the estimated *DIE* temperature Tcore from the CPU Hardware monitor.

Coretemp is probably reporting the cooler T_CASE estimated temperature which is going to be cooler by around 14C than the actual die temperature because that's where your heatsink actually is.

I think both are estimates -- DTS (digital temp. sensor) values depend on the correct value of T_junction_max or whatever for your particular CPU since I think the DTS reads as a delta from the maximum allowed case temperature.

The thermal diode is analog and has to be calibrated in software to translate the voltage on it to a temperature, but it will end up giving something that's roughly correlated also to the core temperature but in a more absolute sense.

Neither measures the actual case/IHS temperature, though that is just sort of estimated based on what Intel says the usual difference between Tcase and Tjunction is which is IIRC around 14C.

Check the FAQ here
http://www.tomshardware.com/fo...quad-temperature-guide

Originally posted by: igloo15
OK so I am ocing my e2180 right now I have a 1.36volt with about an 3ghz overclock (333x9)

Coretemp reads my temperature as 47c while CPUID Hardware Monitor reads it at 62c while running prime95.

Idle the temps are 22c and 37c respectively.

Which should I believe because if its 47c I think i have alittle more headroom to oc more but if I believe Hardware Monitor than I should stop now.

Which temp reading should I believe?

 

[igloo15]

Thanks a wierd thing just happening while running prime95 I had up the oc slightly to 340fsb and was running prime95 fine. All of sudden the monitor goes black and the post screen comes up. My computer never shut off even for a second it ran full speed through the whole thing. When it booted up again it was right back to where I was before. I am not sure exactly what the cause of this could be any ideas?

 

[QuixoticOne]

Sounds like the screen saver kicking in or the PC going into one of its "sleep" modes.
Check your sleep and screensaver timers.

That or your GPU rebooted and the video BIOS reset the GPU as a fault recovery situation when the GPU crashes... you shouldn't see the MAIN SYSTEM BIOS, but you might see the GPU CARD BIOS screen which is temporarily visible as part of the main system POST sequence.

Check your system EVENT LOG and the GPU control panel log and see if there is any indication of the GPU driver or GPU having problems or high temperatures. Try turning on temperature logging for the GPU too.

Either that or maybe a loose monitor cable or static discharge around the GPU.

 

[Basilisk]

I'm going to piggy-back my Qs about temperatures, here, as this discussion has an element of the same problem I'm seeing:
- I just built a G35-based system with an E2180 -- my first experience with a C2D-class intel (a company I'd abandoned for a decade).
- I've been tweeking the OC parameters -- and coming to loathe intel chipsets when using DDR2-8500 RAM as I keep having to re-insert slow, 1.8v RAM to recover
- Running Speedfan with P95 I see Core0/Core1 with matching temps of about 54C, and something labeled "CPU" with a temp that's always about 15C higher than those.

WHAT is this third temperature sensing? If it's the die-temp (WHY is it called "T_CASE"?), how does this differ conceptually from the "Core0/1" temps?? I would generally presume it's the higher temp I should be concerned about, but judging by their NAME I'd have thought the CORE temp was more critical.

Thanks for any insight

 

[Mondoman]

B - check out the link to the temp guide at Tom's Hardware.

 

[Basilisk]

Originally posted by: Mondoman
B - check out the link to the temp guide at Tom's Hardware.

Thanks. That's quite an article, and a starting point for figuring out my system. What had me most confused was that the unrecognized "CPU" [T_CASE] (70C) from SpeedFan was 15C -HIGHER- than the two "core" [T_JUNCTION] temps (55C:54C). Since there were TWO "core" temps, I expected them to be the real CPU temps and the hottest locations -- and that's apparently correct, after reading the THG article.

Apparently I'm facing a wildly -INACCURATE- display of temperatures in SmartFan. If I'm lucky, the displayed T_CASE temp is 20C HIGHER than the correct value, but the only way I'll know is by going through the THG instructions for calibrating offsets.

Thanks again for the pointer, as I was missing the article by searching using the wrong keys.