RealTemp General Discussion

[burebista]

Your sensors are stuck until 40ºC. IMO forget about idle calibration because they'll move only above 40ºC.

But what does means this movement values??? Higher values are better or worst?

It's upside down compared to classic core temperature readings. Higher values of distance to TJMax are better (that's mean that CPU temperature is low).

Core Temperature = TJMax - DTS Data

So TJMax is constant 100 (for your CPU) and higher DTS Data (that's distance to TJMax) mean lower core temperature.

 

[unclewebb]

You can NOT calibrate a CPU that has stuck sensors. Both of your sensors are stuck and do not go lower than 40C.

 

[PsySc0rpi0n]

Your sensors are stuck until 40ºC. IMO forget about idle calibration because they'll move only above 40ºC.

It's upside down compared to classic core temperature readings. Higher values of distance to TJMax are better (that's mean that CPU temperature is low).

Core Temperature = TJMax - DTS Data

So TJMax is constant 100 (for your CPU) and higher DTS Data (that's distance to TJMax) mean lower core temperature.

Can you tell me what is IMO???

I'm sorry but i didn't understand your explanation for the core movement in RealTemp Sensors Test...

So, if i understand, DTS retrieves distance to TJMax, right?

You can NOT calibrate a CPU that has stuck sensors. Both of your sensors are stuck and do not go lower than 40C.

I'm sorry for the stubbornness but i wasn't understanding the purpose of Idle calibration... Now i have understood...

What means the absence of core movement?

 

[burebista]

Can you tell me what is IMO???

Sorry, In My Opinion.

So, if i understand, DTS retrieves distance to TJMax, right?

Yes, that's why you should rely only on this and forget TJMax ambiguous value. Keep distance to TJMax >20-30 and you're safe.

What means the absence of core movement?

That's a stuck sensor. Those weren't design to accurate measure core temperatures but only to trigger thermal throttle when you're close to TJMax so they're (usually) way off (or stuck) at idle/medium load and start to be accurate as your core temperature approach to TJMax.

 

[PsySc0rpi0n]

...

That's a stuck sensor. Those weren't design to accurate measure core temperatures but only to trigger thermal throttle when you're close to TJMax so they're (usually) way off (or stuck) at idle/medium load and start to be accurate as your core temperature approach to TJMax.

Ok. I'm not familiar with shorts for some words...


Some hours ago i've tried to run the sensors test but choosing Large FFTs in Prime95 but i wasn't able to finish the test because of an BSOD. After restart i gave a little more of vCore and tried again to run the sensors test. And when i opened the Sensor Test window of RealTemp to re-run the test i saw there movements in Core 0 and Core 1... (3 and 4 respectively).

With Small FFTs i don't get any Core Movement...

So, how can i interpretate those movements?

 

[unclewebb]

Some movement during the first part of the sensor test is good. Zero movement is a warning that your sensors might be stuck.

The CPU Cool Down Test is more through. It clearly shows that both of your sensors are stuck. That means the temperature of your CPU can go lower than 40C but your sensor will not move at all. RealTemp or any software will continue to report the same 40C temperature for your cores even if you went outside and lived on the North Pole.

 

[PsySc0rpi0n]

Some movement during the first part of the sensor test is good. Zero movement is a warning that your sensors might be stuck.

The CPU Cool Down Test is more through. It clearly shows that both of your sensors are stuck. That means the temperature of your CPU can go lower than 40C but your sensor will not move at all. RealTemp or any software will continue to report the same 40C temperature for your cores even if you went outside and lived on the North Pole.

Ok... But the movement you are talking about is the Core movement or the movement that "Distance to TJMax" shows?

Because i had already had some Core movement as you see in the S.S. (3 and 4 for core0 & core1)... Are these values that i don't know how to interpretate or exactly what they mean...

 

[unclewebb]

The movement test runs an equal load on each core. Each core should change in temperature a similar amount during this test.

If you are running high MHz and high core voltage, then these numbers will be bigger because your CPU heats up more during this test. At low MHz and low core voltage, these numbers will be much smaller because even under full load, a 1600 MHz CPU at 1.1 volts doesn't heat up very much. If your sensors are sticking then your numbers will be lower compared to the same CPU with sensors that don't stick.

The 10 minute test is much better. It makes it very obvious when sensors stop moving as they cool down. At every step of this test there should be movement but once your sensors get down to 40C, they stop moving and report the same temperature, over and over and over again.

Go use your computer and be happy.

Normal sensors move at each step of the test like this:
http://forums.techpowerup.com/showpost.php?p=1061836&postcount=78

 

[PsySc0rpi0n]

This problem of stick sensors is common to every E8x00 or are there many E8x00 that doesn't has this problem??? If i buy another E8x00 am i gonna have the same problem for sure or maybe i don't have it???

 

[burebista]

Is common to 45 nm CPU's. Not every CPU has stuck sensor(s) and are many with OK sensors.
This is my E8400 C0. Pretty decent.

 

[PsySc0rpi0n]

I'll try to purchase another E8x00... Maybe i have better luck this time...

 

[PsySc0rpi0n]

better luck than me , absolutely http://img.techpowerup.org/081120/Capture001166.jpg

But i think you still can calibrate your RealTemp because, at least, one of your sensors doesn't stuck at 40ºC like in my E8400... I think you can work from there because even if the temps aren't gonna be that accurate as if all sensors were working, thay all should have very similar temperature values...

Correct me if i'm wrong...

I wish i had at least one sensor working... That way i could calibrate both sensors based on that working one...

 

[burebista]

I'll try to purchase another E8x00... Maybe i have better luck this time...

And if you get another one with sensors stuck at 60ºC?
Or like a guy in our forums which can't pass above 51ºC whatever he's doing?



My opinion is not to bother, idle temperatures are useless and after 40ºC you have OK readings.

 

[PsySc0rpi0n]

And if you get another one with sensors stuck at 60ºC?
Or like a guy in our forums which can't pass above 51ºC whatever he's doing?



My opinion is not to bother, idle temperatures are useless and after 40ºC you have OK readings.

Well, that is a fact... Better stuck at 40ºC than at 51ºC... By this time i think i have already a buyer to my E8400. If i sell it wish me luck for the next one... lol

Is there any command that i can use in a RealTemp shortcut so that it starts with Administrator previleges instead of a right click on mouse and "Run as Administrator"???

Edited;

Don't need to answer the above question... I had already find out how to do it...

 

[unclewebb]

RealTemp is getting ready for the next official release. It has a new Load meter and calculates the MHz correctly for Core i7.
If you find any bugs let me know so I can get them taken care of before the next release.

http://www.fileden.com/files/2008/3/3/1794507/RealTempBeta.zip

 

[jasondean98]

CPU Cool Down Test

I"ve just run the test on my Q9450, are the results anything to be concerned about?

 

[burebista]

Looks OK to me. 60C in load at 3.2GHz is perfect fine. What cooler do you have and what's your room temperature?
And don't worry about difference between cores this is "by design".

 

[unclewebb]

burebista is right. Nothing wrong with 60C. As long as your computer runs stable at full load and none of your cores are thermal throttling then don't worry about the core temperature.

The sensors for core0 and core2 track each other very well and are likely your most accurate sensors. None of them appear to be getting stuck in this temperature range which is far better than many 45nm Quads.

It's very likely that core1 and core3 have a higher TJMax than core0/core2. Intel actually calls this spec TJ Target. Actual TJMax is typically equal to that spec or somewhat higher but Intel hasn't released any specific details about how much higher TJMax can range from one core to the next. Their opinion is that these sensors shouldn't be used to report absolute temperatures since they're not very accurate or calibrated for that purpose. They were only designed to control thermal throttling and thermal shut down and your sensors look OK for that purpose.

 

[jasondean98]

Edit:

I'm using a true 120 + noctua fan, Antec 900 case and It's summer here in Australia.

 

[burebista]

Perfect heatsink, perfect fan, perfect season in a perfect country so you have a perfect full-load temperature.

 

[jasondean98]

thanks for that! I'll stop worrying now.

 

[Wozzer]

I use realtemp - I think its great.

But can someone explain what TJ Max is?

 

[unclewebb]

TJMax is the maximum junction temperature or core temperature, in theory, that your core will get up to before thermal throttling kicks in. Thermal throttling will drop your CPU multiplier to 6.0 and will kill your performance so you definitely want to keep it below that number.

During testing, thermal throttling usually begins when Distance to TJMax is 2 or 3. Most users with Core 2 will never have to worry about thermal throttling. Core i7 users with the OEM cooler are a different story. Plenty of those users are running full load in the 90C range.

In the Thermal Status area of RealTemp, it will show you if your processor is or has been thermal throttling. OK means there have been no problems. LOG means at least one episode has been logged and HOT means your CPU is presently thermal throtttling.

Once you start thermal throttling it is still possible for your CPU to get hotter if something like your heatsink is loose. The thermal shut down temperature of Core 2 CPUs seems to be about 125C but that number is only documented for the mobile Core 2 chips.

 

[Wozzer]

Thanks - Ive got an Intel E7300.

It runs around 40C when gaming, 35C idle. TJMax 60C. I take it thats alright then

 

[Congo_Toey]

I think something is wrong

I think that there is something wrong with my my temps. I always use your great realtemp tool. I have now moved to the new i7 cpu, and have been using your new version 2.9. But for some reason my idle temps look very high to me even at stock. I am using a koolance CPU water cooling block, and with my E8600 & Q9450 at stock my temps would be around 20 degrees, about 5 or 6 degrees above room temp. For some reason now my idle temps are over 40 degrees even at stock settings. I ran the sensor cool down test at stock setting and over clocked to 3.8GHz. I have attached the screen shots for you. The bit that seems odd to me is the movement values, everyone else seems to get much lower values.