• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

"Core Temp" - CPU Temperature monitoring tool for the K8 series

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Huh?!? I don't see how that could be right
Core_Temp.jpg
 
Update:
Well I've done some reading and investigated largon's results.
It seems that the offset probably applies to the readout given by this sensor. and from what I've seen in some dumps of different chips, that offset could be as small as 1-3C (in either direction) upto over 10C.
I also had a nasty bug where it would not show any temperature below 0 that is now fixed.
I updated to build 85 with offset support and it should now show sub-zero temps, see how it works out for you.
 
Well I guess I was wrong again, and adding Offset detection in most cases FUBARs results. I removed it in build 86.
 
Yep. Definatly.
If you really feel like your avatar suggests, I'd do something about those crazy high temps.
 
Wow if this thing is right my temps are wayyyyy higher than I thought.

In speedfan the onl temp that changed under load went from 28C idle to 36C under load, so I assumed that was my CPU. With this, it shows 52-53C under load!

My CPU is a S754 3200+ Venice at 2475Mhz with a TT Big Typhoon.

How can I be getting that high especially with my HSF? I thought it was mounted ok, I followed all the directions on it and on AS5.
 
Make sure you all update to build 0.9.0.86, and NOT running 0.9.0.85 as it gives FUBAR results.
 
I'm getting wierd results.
My system is in my signature.
At idle, both cores report the same temperature +/- 1 degree and are close to MBM-5 (after adding 10C as an offset in the BIOS -- known problem for early DFI expert MBs).
BUT, when I load prime05 in both cores, one reports 55C and the other soars to 75C! (MBM-5 only goes up to 50C).
When I load 2 copies of super-pi the cooler core reports a little lower (52C) and the other core is _only_ 67C.

How can I have a difference that big when both cores are on the same die so close together? OR, does that explain why core 1 (you call it 2) is the weak core.
 
Yeah, I have a small problem with Core 1 & 2 also - should be core 0 & 1. Eh.



There are some strange results for me, though. Using Prime95 I have the following results (core 0/core 1):
idle - 29/19 (not a surprise since core 0 is the first to fail Prime95)
Prime95/core 0 - 30/25
Prime95/core 1 - 53/29
2x Prime95 - 53/35

It looks like the single runs are reversed or something - weird. But I did double-check the affinity setting and it's right so I don't know what to think. Prime95's affinity setting not as good as first thought?

Also, the VID is showing 1.325v instead of 1.525v - I assume you're reading that off the MB? To go above 1.45v on this MB you have to use a "boost" setting in the BIOS, which adds 0.2v to the VID setting. If that's the real volts as read by the CPU then this chip is a wonder running 2.8 GHz at 1.325v!
 
I didn't notice that VID was wrong (my focis was another calibration on temperature). On my MB, I set it for 1.35v (what's reported), but I set VID Special for 119.2% for an actual Vcpu of 1.609v. I'm guessing that he's not reading the voltage, only the setting registers.
 
davidk21770 said:
I'm getting wierd results.
My system is in my signature.
At idle, both cores report the same temperature +/- 1 degree and are close to MBM-5 (after adding 10C as an offset in the BIOS -- known problem for early DFI expert MBs).
BUT, when I load prime05 in both cores, one reports 55C and the other soars to 75C! (MBM-5 only goes up to 50C).
When I load 2 copies of super-pi the cooler core reports a little lower (52C) and the other core is _only_ 67C.

How can I have a difference that big when both cores are on the same die so close together? OR, does that explain why core 1 (you call it 2) is the weak core.

I would suggest you test what QuietIce has done (#31).
Both idle, load core0\idle core1, idle core0\load core1.

QuietIce said:
Yeah, I have a small problem with Core 1 & 2 also - should be core 0 & 1. Eh.



There are some strange results for me, though. Using Prime95 I have the following results (core 0/core 1):
idle - 29/19 (not a surprise since core 0 is the first to fail Prime95)
Prime95/core 0 - 30/25
Prime95/core 1 - 53/29
2x Prime95 - 53/35

It looks like the single runs are reversed or something - weird. But I did double-check the affinity setting and it's right so I don't know what to think. Prime95's affinity setting not as good as first thought?

Also, the VID is showing 1.325v instead of 1.525v - I assume you're reading that off the MB? To go above 1.45v on this MB you have to use a "boost" setting in the BIOS, which adds 0.2v to the VID setting. If that's the real volts as read by the CPU then this chip is a wonder running 2.8 GHz at 1.325v!

Well Core Temp doesn't use affinity to select a core, its done by writing to special registers in the CPU's IMC.

Regarding VID, its read from the CPU itself, thus on the CPU the vid setting is 1.325v, but the mobo adds the extra 0.2v via the boost option.
 
Neat program! I found something interesting with it too- core 1 (core 0 in SP2004 I am assuming) always runs a little hotter than the other core. This is also the core that has consistently failed prime first while I have been doing stress tests. I know a lot of dual core owners always seem to have a "weak" core, now it looks like this phenomenon may be a part of it.
 
out of: asus probe, everest home edition, bios, speed fan, AND mbm....this program reports 5C lower than all of those. all of the other ones seem to be exactly the same. i dont know what this means, so i'll probably stick with asus probe even though this is prety sweet.

by the way, this reports both cores within 1C or 2C of each other at full load, did i luck out with my ihs? ha.
 
The Coolest said:
Well Core Temp doesn't use affinity to select a core, its done by writing to special registers in the CPU's IMC.

Regarding VID, its read from the CPU itself, thus on the CPU the vid setting is 1.325v, but the mobo adds the extra 0.2v via the boost option.
It is a great program - definitely pointing out the limiter.
:thup: :thup: :thup:
 
The Coolest said:
I would suggest you test what QuietIce has done (#31).
Both idle, load core0\idle core1, idle core0\load core1.
Idle: both 34C

Prime 0: 51/48
Prime 1: 74/51
Prime Both: 75/55

SuperPi Single: 60/50
SuperPi Dual: 65/53
Killed 1st SuperPi: 60/50

(Only about 30-60 seconds each -- not long runs).

4800+
V-CPU 1.609V
2761 MHz (Was stable at 2808, but I've needed to reduce this -- why I'm really interested in the accuracy of the readings)
Danger Den Liquid Cooling

NOW -- what does this mean? That core 1 is running big time hot? And, why doesn't this agree with the BIOS which is supposedly using temperature diodes on the CPU also?

AND -- is there any way to balance? Ie., is it the heat spreader or just that one core runs really hot and there's nothing to be done?

The Coolest said:
Well Core Temp doesn't use affinity to select a core, its done by writing to special registers in the CPU's IMC.

Regarding VID, its read from the CPU itself, thus on the CPU the vid setting is 1.325v, but the mobo adds the extra 0.2v via the boost option.
So, since I set my VID to 1.35V, you show 1.35V. But since you can't see the VID Special of 119.2%, you don't know that it's increased by 19.2%?
 
davidk21770 said:
Idle: both 34C

Prime 0: 51/48
Prime 1: 74/51
Prime Both: 75/55

SuperPi Single: 60/50
SuperPi Dual: 65/53
Killed 1st SuperPi: 60/50

(Only about 30-60 seconds each -- not long runs).

4800+
V-CPU 1.609V
2761 MHz (Was stable at 2808, but I've needed to reduce this -- why I'm really interested in the accuracy of the readings)
Danger Den Liquid Cooling

NOW -- what does this mean? That core 1 is running big time hot? And, why doesn't this agree with the BIOS which is supposedly using temperature diodes on the CPU also?

AND -- is there any way to balance? Ie., is it the heat spreader or just that one core runs really hot and there's nothing to be done?

So, since I set my VID to 1.35V, you show 1.35V. But since you can't see the VID Special of 119.2%, you don't know that it's increased by 19.2%?
Read the two last quotes by me in this post:
http://www.xtremesystems.org/forums/showpost.php?p=1561653&postcount=102

The temperature difference could happen because of a bad mount or poor contact with the IHS.

And you're correct about the VID. That is the reason it says VID and not VCore :)
 
After sleeping on this dilemma and now looking at davidk21770's results I have to wonder if there's something else going on here. I can understand results like his - on Prime 0 BOTH his temps went up. But when running Prime 0 on mine only the core 1 temp goes up, which is exactly the opposite of what should happen (I even checked with Task Manager to make sure I wasn't loosing my mind).

When running Prime 0 I can find no logical explanation for core 1 temps to rise while core 0 remains the same. Has anyone seen a case when core 1 reads higher than core 0 on an AMD?
 
i'm only having an issue with the VID:
huh.JPG


everything is set to 1.40 (none of that % over crap) but this is an old revision DFI ultra-d and its got that funky issue with the voltage to begin with (it being lower than accual set) but 1.425 is quite a bit off

i was wondering, whats TDP?
 
Back