Discussion in 'RealTemp' started by C4rnage, Feb 25, 2012.
CPU is a i5 2500k on Asus P8Z68-V motherboard
Did you run the sensor test?
you mean during that reading or if i did run sensor test someday?
erocker means did you run the sensor test in general? basically the sensor test heats up the cpu and then cools it down measuring the difference over ten minutes untill cooldown, something along they lines anyways, its basically like calibration for the program
heres my 2500K stock on a megahalem after the sensor test in what i would call a general everyday temperature room (just nice not to cold not to warm sorry dont have a thermometer )
i really dont remember if i did sensor test on this cpu or not, but i did after erocker post
how did that go? still the same?
its normal for while,
You might want to try running another monitoring app while RealTemp is running. If they both report 2C and you are not using any extraordinary cooling, my guess would be bad sensors.
Try running RealTemp with the Log File option selected and post that if it shows anything interesting. Copy and paste it into www.pastebin.com and then post a link so I can have a look.
sorry to bump this old thread but it happened again, and this time i had more programs monitoring too, but unfortunately i forgot to turn log on realtemp on.
Given that this problem happened on the same 2 cores, I have to assume that those 2 core temperature sensors have a problem. Intel never intended these sensors to be used for accurate core temperature reporting, especially when the CPU is lightly loaded or idle. If other monitoring programs never report this happening then you will have to use one of them instead.
where is log file located
The log file is usually in the directory where RealTemp is running unless you have told RealTemp to write the log file somewhere else. Right mouse click on RealTemp for the option to relocate your log file.
You might want to try running ThrottleStop for logging purposes. You can use the More Data option in it to gather temperature data 8 times per second. It might help you find this problem quicker.
@unclewebb > Have you noticed that the time stamps of the minimum temps are always the same on the two cores that are reporting low temps; but different from the two that are reporting, seemingly normal temps.
Seems to be an odd behavior.
Just an observation.
95Viper: I did notice the time stamps being the same on both cores but I am not sure what that means.
That chunk of RealTemp code hasn't really changed that much in the last few years and I haven't heard anyone else report this odd behavior so at the moment I am blaming the sensors.
C4rnage: I noticed that you are overclocking your 2500K. What sort of stability testing have you done? Is your computer Prime95 Small FFTs and Blend stable for at least a couple of hours?
Min temps on my quad are fine.
I'm a little miffed that core1 is sucking all the cooling away from core4 though. lol
(core1 is reporting lower than ambient)
yes, i am running my 2500k on 4.3ghz my only problem so far with it is that i need too much voltage, based on other peoples oc to that clock. using 1.36V.
and yea i did run prime95 for some hours and other cpu stress programs too. everything fine except for too much voltage imo.
also i was running realtemp from a shortcut and it didnt log.
I am inclined to believe the same.
Looks as if at low temps the sensors are not accurate or giving erroneous data(as, I have read somewhere that Intel only uses the sensors for processor protection from an over-limit (TDP) condition, not actually temp monitoring through a complete range of temps.)
Anyway, it looks to be ok after the app gets under way.
I would be curious; if, he were to open a second instance of realtemp, would the min temps be correct or give the same results. Actually, it should vary, does every time I try it.
Not meaning to steal the thread, but since Kreij made the statement...
With cooler and board you are using, you may want to check and see, if, the cooler pipes or bracket is resting on the capacitors that surround the processor socket.
I had the same reading when I used a board and cooler of similar layout.
And, the cooler was not sitting flush on the CPU, for that reason.
Just a thought there.
Kreij: 45nm Core 2 based sensors like your QX9650 has get the award for being Intel's worst temperature sensors. Core sensors reading way too low or getting stuck at 40C or 50C and not being able to go any lower were a couple of common issues with these sensors.
If you have unzipped RealTemp into a folder and you have write access to where you located that folder and you don't have the log file open in Notepad or Wordpad preventing new data from being written to the log file, it should work fine. Make sure you also checked Log File in the Settings window and clicked on OK.
A log file would be interesting because it would show if this minimum temperature is a random minimum for 1 second or if it lasts longer than that. Intel's sensors are far from being accurate temperature monitoring devices but I do find they are usually consistent. Random low temperature drop outs are not very common. This is the first I have heard of it.
I know, unc, when I give the proc a little grief all the temps even out nicely and then drop to their goffy idle states, so no problems. I run it at stock, but I must admit I can push my little QX and it behaves quite well with my air cooler.
Separate names with a comma.