DFI Expert CPU temp sensor reporting incorrectly?

[dinos22]

Since the Opteron isn't officially supported by DFI on the nF4 motherboards, you guys are operating outside of the norms.

 

When you're playing with processors that aren't supported you really can't complain about a problem with temp reporting.

i guess we'll just have to wait for that one guy that uses A64 to post his findings hehehehe

[Keekles]

i guess we'll just have to wait for that one guy that uses A64 to post his findings hehehehe

 

 

I have a x2 4400.

Cooling is Xp-90c

 

Running @ 255, 9/10, x2.5

 

Idle temp: 33

Load temp 43

 

Edit: 1.55 +110% Hovering at 1.68

Chipset at 1.7

 

Im comming over to AMD from intel (my last proc ran at 55-60, nicely), so I have no idea about the ranges of these temps.

 

Before OC, even under full load, I did not go over 28*

[dinos22]

here is my CPU code

 

 

I have a x2 4400.

Cooling is Xp-90c

 

Running @ 255, 9/10, x2.5

 

Idle temp: 33

Load temp 43

 

 

 

Im comming over to AMD from intel (my last proc ran at 55-60, nicely), so I have no idea about the ranges of these temps.

 

Before OC, even under full load, I did not go over 28*

what voltage are you using with yours

 

you can calculate power output by this formula

 

(Current vcore/stock vcore)*(Current MHz/stock MHz)*reated stock power output = Total Output

 

My Opteron running at 2.6GHz with 1.45 is putting out 174.45Watts [(1.45/1.32)*(2600/1800)*110W=174.45W]..........that is bound to get crazy hot on air cooling..........obviously ambient temps will affect your temps BUT there is no way in hell you can idle at almost ambient temps with a CPU putting out 174.45 Watts of power...........you'd be on at least 10C over ambient temps......now you can also calculate the thermal resistance of the HSF which would further prove unrealistic and would offer more relativity and substance to these claims....[EDIT]

 

i will in fact take this a notch further

 

 

Here is the thermal resistance formula:

cpu delta = (cpu load) -(room ambient)

 

thermal resistance = (cpu delta) / (total cpu watts)

 

in my case this means

 

12 = 34 - 22

0.06878 = 12 / 174.45

 

 

Hence my thermal resistance is 0.06878............now for anyone that knows anything about air/water cooling they would agree with me that there is no way in hell that AIR can have such low thermal resistance value.......neither can water cooling.......................for an air cooler like mine i would have super excellent results with thermal resistance of 0.18C/W so if you use that value in the fomula you can get 31.40C as the new cpu delta value which roughly means it's pretty toasty in there and my REAL load temp is 53C and not 34C it's reporting

[dinos22]

what's wrong with this photo

 

[Capt Proton]

Dinos22. Where do you have the temperature probe mounted, and how is it secured?

[lancedefrance]

isnt this just a room thermometer? I think he means room temp 28 and CPU 21 but i might be wrong

[dinos22]

isnt this just a room thermometer? I think he means room temp 28 and CPU 21 but i might be wrong

correct

[M_M]

you can calculate power output by this formula

 

(Current vcore/stock vcore)*(Current MHz/stock MHz)*reated stock power output = Total Output

 

My Opteron running at 2.6GHz with 1.45 is putting out 174.45Watts [(1.45/1.32)*(2600/1800)*110W=174.45W]..........

(

Sorry, but I cannot agree with the part stated above!

 

First, 110W is rated MAXIMUM dissipation for whole X2 and dual core opty family, probably correct only for worst case scenario, i.e. X2-4800 @ 2400MHz @ 1.3V.

 

Since P(power)=(voltage^2)/R(resistance), where we can assume that the resistance (seen by the power supply, i.e. voltage source) is constant.

 

So I would rather say (for X2 and dual core opti):

(Current Vcore/1.3V)^2*(Current MHz/2400 MHz)*110W = Total maximum power Output

 

In your case (1.45/1.3)^2*(2600/2400)*110W=148.25W maximum

 

You can see this yourself. Increase the CPU voltage by 15%, put the rig unter heavy load and note the CPU temperature. After that put the voltage back but now increase the clock by 15%, put the rig under heavy load and note the CPU temperature. I can promise you that you will get higher tempereature when only the voltage is increased.

[red930]

what's wrong with this photo

Too much glare from the flash! LOL

[ixtapalapaquetl]

Sorry, but I cannot agree with the part stated above!...

 

In your case (1.45/1.3)^2*(2600/2400)*110W=148.25W maximum...

I am pretty sure that is what he meant. This is the standard equation used. Clearly the voltage is more important than the frequency, and clearly that factor should be squared as a result.

[dinos22]

vozdra belogradjane

 

that's a standard formula used..........ali mozda u srbiji ne vjereruju u zapadnjacke formule.......heheh

 

anyways...........i think we can certainly conclude ASIDE from formulas that the temps are reporting incorrectly

 

 

 

 

Too much glare from the flash! LOL

Looks like Angry_Games's charm is rubbing off on ya ....................

[lorenzox]

Helllo,

 

just to show that these DFI boards doesn't offically support Opteron ---> look at my temp in idle :

Idle temp in celsius, 10 min after a SuperPI 32M successfully completed --- very low !!

 

I have just a question... why some boards report quite non-accurate temps and some other boards report almost correct temps ? :confused:

 

Regards,

lorenzox