E8400 idling at 48c??

S

skyline889

[H]ard|Gawd
Joined
Oct 15, 2007
Messages
1,368
I have my E8400 at 3ghz under volted to 1.05v and Realtemp (Tjmax 100c) reports that my processor is idling at 48c. I have a lapped TRUE with a 1500rpm fan so I have no clue why my temps are so high. I tried re-seating the processor today and re-applying AS5 but the temps remained the same. Did I just get a dud CPU? I would not have spent the extra money had I know it would've run hotter than my E6600. :mad: My E6600 at home would run a 35c idle at 3.6ghz so this is really pissing me off. I had the same high temp problem with the Q6600 SFF build I used to have so it seems like Intel really doesn't give much justification to buy new...
 
Do the temps change if you apply more rpms to the fan or if that fan is capped can you add a more powerful one merely for temp testing? Also what are temps when you crank her up to say 1.3v and approach 4Ghz which I'm sure that cpu can?
Also remember AS5 takes over a week to cure. Now admittedly during that week you should not drop 10ºC in temps though.
Does the hs feel warm?
Software temp monitoring is always going to be a compromise at best.
 
45nm CPUs frequently get stuck at lower temperatures and Intel says this can happen at anything below 50C.
I definitely wouldn't lose too much sleep over your reported temps.
 
Maybe your motherboard needs a bios update? Could be reading temps funny. I had that problem with my Abit IP35Pro.
 
CPU peaks at 51c at full load so I don't think the sensor is stuck? I can't adjust fan speed since the P5Q Pro is retarded and won't let you change it but when I overclocked it to 4ghz at 1.35v, the idle temps went up to 58c and load temps were over 60c iirc. At this point I have no clue what's going on so I think I might just ditch this chip and go with a used E6600 or 7200.
 
Umm duh, 58c idle is not normal for an E8400 and a lapped TRUE with only 1.35v. Sorry but why do you have to come off like such a douche? He asked what temps were like when overclocked, I wrote an answer.
 
Idle temps are nowhere near accurate on the 45nM chips and Intel has acknowledged that the first accurate temp is when you reach TjMax.
So load temps are much nearer correct, idle temps are way out.
 
lowteckh said:
You said the temps were 51c load at stock, then become confused when the temp is higher when it's overclocked. So yea I'm gonna give you a duh answer. If you even bothered to look around the web or even here for e84 temps at that clock, voltage and heatsink you'll see that it is normal. Thx for doing your HW.

http://hardforum.com/showthread.php?t=1314774&highlight=e8400+true
Click to expand...
I've researched my information. A lot of people complained about stuck sensors, I don't believe I have one. I'm not confused that the temps went higher; someone asked what my temps were at 4ghz and 1.3v, I posted my results. At no point did I ever state that the temps were not what I was expecting. My post was simply a response to a question that was posed by someone who was trying to help me figure this out. The confused part was only in reference to why my chip has such a high idle, regardless of OC amount. This is not my first time overclocking, obviously temps go up as you OC. Act your age man, and read before you lip off.

As for your link, I read that thread before I posted. Different cooler, completely different load temps, not the same situation.
 
Read before YOU lip off someone trying to solve your problem that you're so oblivious to. What is so different about Nenu's temps? Idle temps don't matter a drop with the e8s and his load temp is way similar to what you are putting out.
http://hardforum.com/showpost.php?p=1032605144&postcount=2
Nenu said:
With Realtemp figures of 43C idle and 65C load at 1.4V, you temps are fine.
I am running an E8400 @ 4.13GHz, 43C idle and around 67C load at 1.4V set in BIOS
TRUE 120 Heatsink.
Click to expand...
skyline889 said:
when I overclocked it to 4ghz at 1.35v, the idle temps went up to 58c and load temps were over 60c iirc.
Click to expand...

Why I even bother is beyond me.
 
Disregarding the fact that those temps were posted back when everybody believed the TjMax for E8400s was 95c and not 100c, the results are completely different. The OP of that thread yes, was getting load temps of 65c, but had temp movement of over 20c. I had movement of about 3-4c between load and idle in both non-oc and oc'd form.

As for Nenu, he was getting the same kind of temperature movement as kooty and had 10c lower idle temps (58c compared to 48c, adjusted for 95-100 TjMax change) at a higher oc with more voltage and a non-lapped TRUE (Good for ~5c drop in temp).

Obviously something's wrong since I'm only getting a temp increase of 3-4c between idle and load but I don't believe it's the result of a stuck sensor since the instances I've read of don't have the idle and load temps moving much at all between different OC speeds, whereas mine jumped 10c at the different idles and loads. It very well could be a stuck sensor but at this point, I'm really not sure. This is all I was confused about, not the temps rising as I increased the OC. Obviously that's going to occur as you increase voltage and speed.
 
Idle temps can vary for the e84s. I've had a C0 and E0 on the same cooler and both posted different idle temps (C0 was stuck at like ~50, changed 3 different coolers NT06, Enzotech Ultra, S1283. E0 was reading ~40-50, even less I think) but similar load temps. Seriously, don't even bother looking at them. The load temp is the only thing you should be looking at, the only thing, and yours fits just right in there. 1.4v and <70c is just about it for air cooling for this chip.
 
lowteckh said:
Idle temps can vary for the e84s. I've had a C0 and E0 on the same cooler and both posted different idle temps (C0 was stuck at like ~50, changed 3 different coolers NT06, Enzotech Ultra, S1283. E0 was reading ~40-50, even less I think) but similar load temps. Seriously, don't even bother looking at them. The load temp is the only thing you should be looking at, the only thing, and yours fits just right in there. 1.4v and <70c is just about it for air cooling for this chip.
Click to expand...
Seriously dude, that's all you needed to say. What had me pegged was there only being a few c degree change between idle and load. Thanks for the advice though, I knew temp reporting was a problem with the Wolfdales but I hadn't read about it in a while so I figured Intel had gotten it worked out by now.
 
skyline889 said:
As for Nenu, he was getting the same kind of temperature movement as kooty and had 10c lower idle temps (58c compared to 48c, adjusted for 95-100 TjMax change) at a higher oc with more voltage and a non-lapped TRUE (Good for ~5c drop in temp).
Click to expand...

To be clear, adjusting TjMax has no effect on the actual measured temp, only the delta to TjMax will change.
You are trying to resolve an impossible situation here.
Verify load temps, not idle.
Load temps are more likely to represent something useful.
 
Nenu said:
To be clear, adjusting TjMax has no effect on the actual measured temp, only the delta to TjMax will change.
You are trying to resolve an impossible situation here.
Verify load temps, not idle.
Load temps are more likely to represent something useful.
Click to expand...
Not 100% but I believe it's actually the opposite of this. Reported core temps in Real Temp jump 5c when going from a 95c TjMax to a 100c TjMax. Distance to TjMax (Delta to TjMax) remains the same for me.
 
The e72 has a more messed up sensor from what I just read if you're seriously considering it cuz of the idle temp.
 
skyline889 said:
Not 100% but I believe it's actually the opposite of this. Reported core temps in Real Temp jump 5c when going from a 95c TjMax to a 100c TjMax. Distance to TjMax (Delta to TjMax) remains the same for me.
Click to expand...

lol, I am backass and have done that before over these temps.
My bad, you are correct.
Must be those aluminium seasoned frog cakes I've been eating, bringing on Alzheimers.
 
Lol no worries dude, its 430am here so I'm mixing myself up a lot too.

__

I know I should just forget about idle temps but it's really bothering me so I think I might jump ship and go back to 65nm. Even at 3ghz, the E6600 did everything I needed fine and it'll run even cooler up here since its like 40f compared to 85f outside (Oregon vs Hawaii) so I think this is the way to go for me. I'll save myself a few bucks too with E6600 prices at just 80 bucks now!
 
Do the stuck sensor test on Realtemp to see if you have any. Mine both say unstuck, but my idle temps are still high at around 40C, so YMMV.
 
Intel officially states that the 45nm sensors in their Atom CPUs when at 50C are accurate to plus or minus 10C. That's a pretty wide range.

There are 3 major problems. TjMax might have an official Intel spec of 100C for your processor but they haven't stated how much error is in that number so it's a meaningless spec. If the error is plus or minus 5C which is typical, then your actual TjMax might be 95C, 100C or 105C.

Add in some slope error where the sensor is moving at a different rate than changes in temperature and the possibility of them sticking as the sensor tries to go under 50C and you end up with a big mess. If you're not able to calibrate these sensors then they're pretty much useless for accurate temperatures.

Did you try a program like SpeedFan which also shows you temperature data but from a different sensor. When you have bad core sensors, the sensor simply labeled as CPU in SpeedFan might be the better way to go. It's not 100% accurate either if you don't calibrate it but it usually has a wider range and doesn't have as many sticking issues in the normal temperature range.

If I'm running at 4 GHz and I'm stable then I don't worry too much about what core temperature is reported. These sensors are flawed but the 45nm CPUs work great.
 
unclewebb said:
Intel officially states that the 45nm sensors in their Atom CPUs when at 50C are accurate to plus or minus 10C. That's a pretty wide range.

There are 3 major problems. TjMax might have an official Intel spec of 100C for your processor but they haven't stated how much error is in that number so it's a meaningless spec. If the error is plus or minus 5C which is typical, then your actual TjMax might be 95C, 100C or 105C.

Add in some slope error where the sensor is moving at a different rate than changes in temperature and the possibility of them sticking as the sensor tries to go under 50C and you end up with a big mess. If you're not able to calibrate these sensors then they're pretty much useless for accurate temperatures.

Did you try a program like SpeedFan which also shows you temperature data but from a different sensor. When you have bad core sensors, the sensor simply labeled as CPU in SpeedFan might be the better way to go. It's not 100% accurate either if you don't calibrate it but it usually has a wider range and doesn't have as many sticking issues in the normal temperature range.

If I'm running at 4 GHz and I'm stable then I don't worry too much about what core temperature is reported. These sensors are flawed but the 45nm CPUs work great.
Click to expand...

Intel has released TJmax information to the public, but you're right, I didn't see any quote on accuracy.

http://www.hardforum.com/showthread.php?t=1338140
http://www.hardforum.com/showthread.php?t=1336257

Personally, I would just set up Realtemp properly and monitor my load temps. The idle temps should not worry you if your load temps are in order. IIRC, both Speedfan and Everest will show you those additional temp sensors and may provide some more insight.
 
What Intel released to the public at IDF was more of a PR exercise than anything else. So much information has been left out that software and programmers are still left making assumptions and guessing. Their most recent release where they are now using the term TJ Target has just caused more confusion. There are many 65nm processors out there where TjMax and the newly released TJ Target spec are two different numbers. It's hard to calculate an accurate absolute temperature number if TjMax isn't accurate.

Any company like Intel is going to have huge volumes of information about these sensors but after two years of begging, they've only released the bare minimum amount of information to keep people happy.
 
You must log in or register to reply here.
Back
Top