Not Sure What To Make Of This...

[Excidium00]

Well, I just installed my Zerotherm FZ120 to see how my temps would be on my E8400 at 3.6ghz and I'm seeing something strange on my CoreTemp. It's reading 49C on core #0 and 42C on core #1 at pretty close to idle, like 3% load, as well as a very similar difference while at load. Strange thing is that if core #1 is even at higher load, it has lower temps than core #0. I'm not sure what to make of this, really. Also, RealTemp and CoreTemp are registering temps 10C off each other, so I don't know what is going on.

[Syngensmyth]

Google core temp difference or some such. You will find it is not uncommon. Could be uneven heat sink contact or uneven thermal paste or uneven pressure somehow. The internal sensors may be off somehow. Try different programs for reading the sensors to be sure one program is not reading wrong.

 

Some people "lap" the contact areas to be certain of even contact. Maybe not worth it unless your temps are too high or you are desiring a higher clock.

[Excidium00]

Everest and CoreTemp give me pretty close temps to each other. RealTemp is giving me 10C below those 2 and Speedfan gives on average 5C below what Everest and CoreTemp read. Nothing wants to make up its mind. I think I'm gonna disconnect everything again so I can check the thermal compound and the seating, but they looked pretty good yesterday. I do want a higher clock, around 3.8, but I think my current ambient temps might be throwing everything off.

Edited July 31, 2008 by Excidium00

[Syngensmyth]

Here is an interesting thread right now on lapping.

http://forums.overclockersclub.com/index.p...=155393&hl=

 

I hate speed fan. Just an opinion.

Edited July 31, 2008 by Syngensmyth

[dr_bowtie]

I dont use anything but Everest for temp monitoring....the free trial will work fine...

[SebringTech]

I would recommend you try lapping the cpu and heat sink if you are comfortable with doing so. I did and saw my core temperatures get a few degrees closer together. Then again the IHS on my Q6600 was quite uneven.

[Excidium00]

Sounds like a plan, I have an old P4 I can play around with before i go for it on my e8400. Now I just gotta find lots of sandpaper

[s0rd3z]

The thing with real temp is that it is using different tjmax values to determine the max temp threshold, so some values are different compared to core temp and other temp monitoring software.

 

Here is a link to a forum where a lot of discussion during the development of realtemp went on. Why they made the changes, and there are posts that explain how to check for stuck sensors or other things in the prog.

 

http://www.xtremesystems.org/forums/showthread.php?t=179044

 

It's a ton of pages, and you'll have to sift through a lot of somewhat redundant info. But I did a lot of research before I decided that realtemp seems, after checking all the other utilities, the most accurate for intel based cpu's at this time.

 

Now that doesn't negate the other temp monitoring software, it's just the conclusion I came to personally after a day or two of overclocking and testing with a lot of different temp monitoring software.

 

I would still do the lapping and other suggestions cause that seems a more likely cause, and keep using other temp software of course. RealTemp is by no means the be all end all, and it can be confusing cause most other temp monitoring software is using the same values to determine these things. I just wanted to throw out some info as to why you get that 10c difference compared to other progs consistently. RealTemp is the only one that I know of that is using there own values that they came up with after a lot of voltage and temp testing, which is mostly documented in that thread I linked.

Edited July 31, 2008 by s0rd3z