This is Patently Ridiculous...

Toggle sidebar Toggle sidebar
B

blankets

Member
Apr 8, 2010
31
0
0
So i just built a new system, phenom x3 720 BE with a hyper 212+ cooler. i unlocked the 4th core, tested stability, and then decided i should probably find out the temperature offset between my TMPIN0 and TMPIN1 and the actual CPU temps because after you unlock it, CPU temps for the cores no longer work

i used HWMonitor for temp monitoring.

When i re-locked the core and got the temp monitors again, i got some Very interesting results...

proof.png


wtf? 36 load? the picture doesn't show it but i have OCed it to 3.5 Ghz with a core voltage of 1.45 V

to make it worse, when i let it idle, the temps went down to 23 C. I don't live in antarctica. it is room temperature here. This is Impossible.


i presume this is something wrong with HWMOnitor?
 
B

Ben90

Platinum Member
Jun 14, 2009
2,866
3
0
Ive known HWmonitor to be very off. Try something that tells you the TJmax directly.
 
BD231

BD231

Lifer
Feb 26, 2001
10,568
138
106
Could be right actually, when I first bought an x3 and wasn't overclocking I had idle temps of 16 to 18c on cool days. My x3 idles at 20 to 22c @ 3.73ghz, that's just cool and quiet downclocking the cpu. At full load I hit 46c, 48c if its hot out. As for your load temps, you're not really pushing your voltage that high. 1.45v is more like 1.425 actual on a lot of boards so this might be contributing to the lower temps. Its also a fresh build, temps might go up over time.
 
A

Athadeus

Senior member
Feb 29, 2004
587
0
71
That's not very long into the test, and prime95 is not the most stressful either. I'm getting one of those 720s delivered really soon and am planning to stock cool it so I like hearing this if it ends up being accurate.
 
B

blankets

Member
Apr 8, 2010
31
0
0
Athadeus said:
That's not very long into the test, and prime95 is not the most stressful either. I'm getting one of those 720s delivered really soon and am planning to stock cool it so I like hearing this if it ends up being accurate.
Click to expand...

I took that picture a bit early into the test. i ran it for an hour and the temps never fluctuated from 36 C.
 
B

blankets

Member
Apr 8, 2010
31
0
0
BD231 said:
Could be right actually, when I first bought an x3 and wasn't overclocking I had idle temps of 16 to 18c on cool days. My x3 idles at 20 to 22c @ 3.73ghz, that's just cool and quiet downclocking the cpu. At full load I hit 46c, 48c if its hot out. As for your load temps, you're not really pushing your voltage that high. 1.45v is more like 1.425 actual on a lot of boards so this might be contributing to the lower temps. Its also a fresh build, temps might go up over time.
Click to expand...

on a semi-related note to that, i'm overclocking it with 4 cores unlocked. do you think 1.5 V is too high in terms of cpu degradation?

i know amd's site says that the x3 720's max safe voltage is like 1.425 but other forums have posted that anything under 1.55 v should be safe for phenom 2 x4's. thoughts?
 
MJinZ

MJinZ

Diamond Member
Nov 4, 2009
8,192
0
0
blankets said:
So i just built a new system, phenom x3 720 BE with a hyper 212+ cooler. i unlocked the 4th core, tested stability, and then decided i should probably find out the temperature offset between my TMPIN0 and TMPIN1 and the actual CPU temps because after you unlock it, CPU temps for the cores no longer work

i used HWMonitor for temp monitoring.

When i re-locked the core and got the temp monitors again, i got some Very interesting results...

proof.png


wtf? 36 load? the picture doesn't show it but i have OCed it to 3.5 Ghz with a core voltage of 1.45 V

to make it worse, when i let it idle, the temps went down to 23 C. I don't live in antarctica. it is room temperature here. This is Impossible.


i presume this is something wrong with HWMOnitor?
Click to expand...

Correct - HWMonitor is not useful for core temperatures.
 
dzoner

dzoner

Banned
Feb 21, 2010
114
0
0
blankets said:
So i just built a new system, phenom x3 720 BE with a hyper 212+ cooler. i unlocked the 4th core, tested stability, and then decided i should probably find out the temperature offset between my TMPIN0 and TMPIN1 and the actual CPU temps because after you unlock it, CPU temps for the cores no longer work

i used HWMonitor for temp monitoring.

When i re-locked the core and got the temp monitors again, i got some Very interesting results...

wtf? 36 load? the picture doesn't show it but i have OCed it to 3.5 Ghz with a core voltage of 1.45 V

to make it worse, when i let it idle, the temps went down to 23 C. I don't live in antarctica. it is room temperature here. This is Impossible.


i presume this is something wrong with HWMOnitor?
Click to expand...

I have a 720BE too and noticed the same thing with HW monitor. I have a Xigmatech S1283 cooler, which is roughly comparable to yours, in a HAF32 case.

My thinking is that's WAY better cooling than stock and, at minimum, more than sufficent to cool even an unlocked overclocked 720, especially if your case has good air flow. So I set a fairly cool CPU temperature cutoff point (80C) in BIOS in case the fan stops for some reason and so the CPU doesn't go to it's (much higher) internal temp. cutoff point and don't give it much more thought.

I used to do video encoding on my main computer and had the 720 on four cores at 3.5 ghz, but now I have a 720BE in a second computer oc'ed to 3.5ghz, though only with three cores active, and I've done continuous days of video encoding, which kicks cpu utilization near 100% on all three cores and never had a problem and that on a Xigmatech Rifle 90mm cooler.
 
Last edited:
D

deimos3428

Senior member
Mar 6, 2009
697
0
0
There's no way that the individual cores under load are going to be less than TMPIN0. The 36C figure is incorrect, and if I were to guess it's probably closer to 48-50C.
 
BD231

BD231

Lifer
Feb 26, 2001
10,568
138
106
blankets said:
on a semi-related note to that, i'm overclocking it with 4 cores unlocked. do you think 1.5 V is too high in terms of cpu degradation?

i know amd's site says that the x3 720's max safe voltage is like 1.425 but other forums have posted that anything under 1.55 v should be safe for phenom 2 x4's. thoughts?
Click to expand...

AMD stated 1.5v is safe on the SOI process at one of their overclocking events, you certainly don't have to worry about voltage like you do on Intel chips. Of course, most would tell you more voltage will shorten the lifespan to some extent; but in terms of consequence I highly doubt you'll be running that chip by the time you notice any.

Worst case, you take a few years off the top of the chips lifespan. With cool and quiet my chip doesn't really see full voltage half the time it's on anyway, so if you plan on browsing the internet as much as you do anything else with your computer the picture should be the same for you.
 
TurtleCrusher

TurtleCrusher

Lifer
Apr 20, 2008
10,161
984
126
Realtemp gives me the most (seemingly) accurate temps for my Q8200. Most likely it's just as good on AMD boards.

If you want to test your CPU/RAM, use Linx. It's the only stability app I use because it find errors really quick.
 
Sylvanas

Sylvanas

Diamond Member
Jan 20, 2004
3,752
0
0
Could be a BIOS issue. Sometimes manufacturers release an updated BIOS to deal with funky readings coming from their boards- I'd check that. If that's the case, no matter what app you use it'll give you the same reading as what the sensor is telling the OS even though it's wrong.
 
B

blankets

Member
Apr 8, 2010
31
0
0
Sylvanas said:
Could be a BIOS issue. Sometimes manufacturers release an updated BIOS to deal with funky readings coming from their boards- I'd check that. If that's the case, no matter what app you use it'll give you the same reading as what the sensor is telling the OS even though it's wrong.
Click to expand...

the bios is fairly recent and i am hesitant to flash my (otherwise) working bios. i guess i'll just keep an eye on the tmpin01
 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom