CoreTemp vs RealTemp :::: E8400

T

TrueChaos99

n00b
Joined
Nov 26, 2007
Messages
10
I'm confused over which one of these two programs is actually reading the temperature correctly for my E8400.

I'm on the stock intel cooler, and coretemp read 45C idle and 65C load, as the maximum. Occasionally one of the cores would be at a different temp, but not usually. Everest reads the same as coretemp. RealTemp on the other hand, reads 10C lower than coretemp. Which one of these is correct? I'd like to believe realtemp, as I want to run 3.6 on the stock cooler -- this gives temps of 70C in CoreTemp and 60C in RealTemp. If Coretemp is right, I'll back off until I can afford a better cooler, but I'm more inclined to believe realtemp. This is because it just doesn't make sense that a processor would hit 65C load under 100% stock conditions. My case is not cramped, and does have good airflow. Which one is correct?
 
thanks. I did try searching, but wasn't coming up with anything. Even pointed google to this subforum with no results.
 
There are threads on every computer forum and in every language on this topic. The internets must be brokens.

Realtemp for 45nm chips.
 
UncleWeb wrote Realtemp because he found discrepancies between the actual temps (using a heat sensor) and the reported temps from temp utils such as Coretemp.
He has included the ability to modify the idle temps if they are out and has knocked off 10C from the reported temps and the max temp (Tjmax) for the 45nM series chips as he found they are too high.
 
I'm running Realtemp on a 65nm e4300 and have complete faith in it. Yep...it's about 10-12 c cooler than Everest and Coretemp.

Awesome as now I can bump up the OC....
 
Realtemp and Coretemp read the same if you adjust them to display the same tjmax. Both programs allow this so use which every you like better. Example, my x3360 running the x264 benchmark hd during the end of one of the 2nd pass runs. ALL THE SCREENSHOTS BELOW WERE TAKEN LIKE THIS (yes, my X3360 runs cool):

samepl0.gif


Now, if I use the same two apps without modifying the coretemp.ini file to drop the tjmax from 105 to 95, coretemp is 10 °C hotter as you can see:

seehb2.gif


Finally, here are all four apps (everest, hwmonitor, realtemp, and coretemp). Note that I did NOT adjust the tjmax in coretemp back to 95, also, the 1 °C discrepancies you see are a function of polling time, in other words, all the apps aren't polling the DTS in the core at the same time:

finallyrj4.gif
 
Now, if I use the same two apps without modifying the coretemp.ini file to drop the tjmax from 105 to 95, coretemp is 10 °C hotter as you can see:
Click to expand...

Might want to double check your images. ;)
 
RealTemp FTW. I'm with GraySky about adjusting Core Temp, though... I like it better.
 
Nenu said:
UncleWeb wrote Realtemp because he found discrepancies between the actual temps (using a heat sensor) and the reported temps from temp utils such as Coretemp.
He has included the ability to modify the idle temps if they are out and has knocked off 10C from the reported temps and the max temp (Tjmax) for the 45nM series chips as he found they are too high.
Click to expand...

I agree 100%, but I still don't understand the ever present popular discussions about the temp differences between Coretemp, RealTemp, SpeedFan or any Intel C2D (duel or quad) cpu temp program. On all three (3) of the boxen I have folding 24/7 I have had the same temperature readings per program. (oh yeah, they're all duel booter's, WinXP/Linux). I'm reminded of the ole' expression "garbage in and garbage out" because all the above mentioned programs read the same TJMax core diodes (I'm pretty sure) and if they don't have the same readings then one of the affore mentioned programs has a different hard coded formula for equating the cpu core temps or isn't using the same hard coded TJMax numbers (ie 85c or 100c) :D

It seems to me to get the most accurate readings from a software program is to freaking calibrate it. It ain't hard. Hell, even uncleWebb explains how to calibrate his program called "CoreTemp" in simple English. He uses small untechnical words for us intelligence challenged (that's PC for dummies) and I have no idea if it's been translated to other languages.

It seems to me if you are highly overclocking your cpu (AMD or Intel), don't have a stuck core diode sensor (I've only read about the Intel penryn chips having this problem) and are extremely worried about your cpu temps then IMHO you should get a calibrated hardware sensor because they're the only way to get close to 100% accuracy. That being said I think most of the software temp programs (I'm partial to RealTemp myself) are close enough for us regular computer people for accuracy concerning the temperature readings of our cpus. Any software programs IMO are only good for reference readings anyway, like to tell you if it's "dust bunny" cleaning time, if you have a HS cooling fan problem, messed up HSF installation, getting into a too "hot" temperature while overclocking or in my case a bad WC'ing pump, etc.

Nothing BTW is 100% accurate (with the exception of God or the top diety of the religion you choose ) :)

These are just my personal opinions and of course could be wrong. Please don't anyone have a "heart attack" if you don't agree. :(


 
graysky said:
Realtemp and Coretemp read the same if you adjust them to display the same tjmax. (yes, my X3360 runs cool):
Click to expand...

I agree that's one cool running CPU :eek:. While I don't have that much headroom on my e4300, I typically have 20c headroom before the "stock" TjMax so I don't feel the need to adjust TjMax. I did do the idle calibration though, and don't really care about the temp differences between the various programs. Especially @ idle..... My system doesn't idle much... :D
 
lol too bad you cant get accurate temps with a phenom 9600 :( core temp says i am @ 32c :D
 
You must log in or register to reply here.
Back
Top