RealTemp 2.5

Haha... This is far more accurate than core temp IF and ONLY IF you calibrate it properly, Actually it's probably better uncalibrated. I notice they are using a 85ºC TjMax for a Q6600.
 
I mentioned at XS that I'm not 100% sure if TjMax=85C or 95C for the B3 Q6600. I've tested my G0 and I know it is TjMax=95C but the original Quad might be 10C lower. I need a user or two to post some idle results at 1600 MHz and 1.10 volts with a B3 compared to their room temperature. Idle and room temp should be about the same if TjMax is correct for the B3. Better yet, you could send me a B3, I'll thoroughly test it and will send you a G0 in return for your troubles. I would think that would be a deal no one could refuse but I have no takers yet. Having a B3 to do some testing with would clear things up.
 
There's always users asking for more! Any suggestions? I'm planning out the next round of features and upgrades for RealTemp.

The only thing I won't add is support for AMD processors. I'll let CoreTemp handle that one. RT is only for Intel Core processors.

Yeah, you gots that right uncleWebb. :D

Thank you a whole bunch for coming over to our forum and replying to some of our posts

I apologize for my above post about the incorrect statements of RealTemp possibly supporting the AMD chips (I haven't used an AMD chip since Intel came out with the C2D's). I corrected it by editing it to prevent any more bogus info. Thanks for supplying your accurate info. (can't get more accurate information about a program then from the author himself) ;)


 
works for me.. not sure if it is better/more accurate than core temp though.

Personally I think RealTemp is without a doubt the best and most accurate software temperature reading program out here ATM. The only thing better IMO is a calibrated hardware sensor machined into the heat spreader on the chip and even with that it's impossible to check the "hot spots" on each core of a multiple core chip.

It's been written in every forum, white paper, computer magazine, etc many, many times that the C2D cpu has what's called a TJMax (TDP envelope, throttle point, better cut down or you'll fry point) for each core, either 85c degrees or 100c degrees.I think all software temperature programs use these figures (85c or 100c, with few exceptions) unless they monitor the case temp or the temp of the whole cpu itself instead of each core like mobo sensor readings ( these numbers are hard coded in the program) The simple reason why they use these numbers is because Intel won't publish the actual numbers. Thanks Intel. (that's why the adding and the subtracting of15 degrees bullcrap, from my understanding the actual temps aren't linear anywhoo) and they can not be changed (unless you happen to be a computer programmer, with all their expensive compilers) they are hard coded in the program. It wasn't until uncleWebb was nice enough to write the RealTemp cpu software temperature monitoring program for the Intel C2D processor that any ordinary computer user could tweak these numbers to the correct amount for their particular cpu (85c to 100c, change them)



Now some of this information may be incomplete, hard to understand or, heaven forbid, even wrong, if it is I apologize. In my book RealTemp is the best software cpu temperature reading program out here at the moment. I give my gratitude to uncleWebb for inventing it.


 
Quick question, if RealTemp hasn't been properly calibrated then will both idle AND load temperatures possibly be inaccurate? Or just idle?
 
Quick question, if RealTemp hasn't been properly calibrated then will both idle AND load temperatures possibly be inaccurate? Or just idle?

I read the temperatures on the Intel C2D cpu were what they call "non linear", in other words the chips distance from it's throttling point is not read evenly from idle to load. (TJMax)

So my answer to you is IMO uncleWebb has made calibrating his RealTemp program so easy a mentally challenged three year old could do it. Not having the temperatures calibrated IMO is an exercise in futility. Hell, you might as well do a "finger test"(gots to be careful, don't want to fry yourself or your machine), flip a coin and pick any arbitrary number out of the sky as your chips temp. (not really, I think RealTemp and Coretemp are both accurate programs out of the box. Not near as accurate as RealTemp is when it's calibrated though)

In all seriousness if the temp program you are using has a TJMax of either 85c or 100c hard coded into the programs code, which, as far as I know all the programs on the market today that read Intel multiple core chips do. (diodes?) Then you can hope the author guessed the correct guess for your chips TJMax. (as far as I know Intel won't publish the TJMax of any of it's core sensors) and your chip temps are accurate

I think the greatest thing about the RealTemp program is, as far as I know, this is the very first time you can adjust your TJMax on the fly and custom fit it for your cpu. :p

If I'm mistaken about anything, please correct me :)

 
Another supporter of RealTemp, thanks UW! Love it, and use it on all my puters.
 
cool! do you think you can make it send an email at a certain temperature? or even shut the system down. would be great for render/folding farms which rely on the room being cool, and sometimes the AC fails when nobody is about.
 
Now if we only had a version that worked so will for AMD!!! Thanks I will check this version out.
 
The updated version soon to appear will let you enter in what ever TjMax you want from 70 to 120 and it won't have to be a multiple of 5 like previously.

The idle calibration factors will also be expanded some more to include one digit after the decimal point. Can't decide on 1 or 2 then you'll be able to use 1.5 or 1.7.

Temps will also be showing up in the Tray area. My daughter isn't happy that her little icon will be disappearing but she understands that I need to keep up with the good features that CoreTemp has. It will be totally configurable so users can choose exactly which cores they want to look at from 0, no cores to all 4 on a Quad or any combination in between.

Without any calibration, RealTemp should be very accurate for load temps as long as your load temps are over 60C. Below 60C, the accuracy of the digital thermal sensors start to decrease and decrease and decrease. In my cool basement I have an E6400 and an E8400 that are both off by about 8C at idle.

I found a couple of users over at XS and they did some testing and TjMax=85C still seems correct for the early B3 Quad processors. Users keep trying to prove me wrong but so far I haven't seen any evidence to change my opinion on that one. It wasn't until the G0 stepping that Intel bumped that up to TjMax=95C.

The early thinking that TjMax must be either 85C or 100C doesn't seem valid for the desktop processors. My opinion is that old tale only applies to the mobile processors but even that isn't really true because I haven't seen any TjMax=85C mobiles. Those two magic numbers might date back to the late Pentium4 era when these sensors were first introduced.

If anyone is still unsure if RealTemp or CoreTemp is telling them the truth then check out some of the testing that is behind RealTemp.
http://www.xtremesystems.org/forums/showpost.php?p=2883315&postcount=573
 
Quick RealTemp (and probably CoreTemp) question.

If I have reason to believe my motherboard sensors might be faulty, like with Speedfan, does RealTemp bypass any motherboard peculiarities and read directly from the CPU?

Robert
 
All of the "Core" temp reading utilities take their information straight from the thermal diodes on the processor cores themselves - the motherboards never have anything to do with the readings whatsoever. CoreTemp, RealTemp, Speedfan, etc... they all read the info directly off the CPUs.
 
where can I find vid in this program?

There's a tiny little nub of a button up next to the processor speed, upper right hand corner. Each click of that brings up different info, I think 3 clicks gets the VID reading.
 
All of the "Core" temp reading utilities take their information straight from the thermal diodes on the processor cores themselves - the motherboards never have anything to do with the readings whatsoever. CoreTemp, RealTemp, Speedfan, etc... they all read the info directly off the CPUs.

Hmmm. Okay. I have a DFI board exhibiting similar behavior to this guy and am trying to figure out my problem. I'm down to either the mobo sensors (if this is possible), or a very bad Thermalright heatsink.

Robert
 
Back
Top