I hear Abit shows a wrong CPU temp, true ?

[ZL1]

Hi

guys Ive heard that Abit will show a wrong CPU temp and I was wondering if thats true
I was kinda hoping it is because Im building a machine for a friend using the abit is7-e and right now the temps are scary : 65deg load / 50 idle


please let me know


Thank you
Dan

 

[RalfHutter]

Yes, it's true. Abit does it on purpose.

 

[ZL1]

Originally posted by: RalfHutter
Yes, it's true. Abit does it on purpose.

Thanks !

 

[Peter]

Well, in fact, according to that thread, they're showing the actual chip temperature, as measured by its internal probe. Yes that gives unusually high readings to people who are used to readings from external sensors under the CPU, inside the CPU socket. However, it's the much more accurate and relevant reading.

The CPU's own thermal probe is very close to its hottest spot, and yes, the temperature there changes very very quickly depending on CPU load. That's the whole point of having it.

Athlon XP processors are specified to operating temperature limits between 80 and 90 degrees Celsius. So people who are getting nervous at 70 need to reeducate themselves ...

 

[ZL1]

Originally posted by: Peter
Well, in fact, according to that thread, they're showing the actual chip temperature, as measured by its internal probe. Yes that gives unusually high readings to people who are used to readings from external sensors under the CPU, inside the CPU socket. However, it's the much more accurate and relevant reading.

The CPU's own thermal probe is very close to its hottest spot, and yes, the temperature there changes very very quickly depending on CPU load. That's the whole point of having it.

Athlon XP processors are specified to operating temperature limits between 80 and 90 degrees Celsius. So people who are getting nervous at 70 need to reeducate themselves ...

Honestly I doubt this and abit hasnt confirmed it so far (page 7)
I'll explain: if they used this more accurate new standard then they wouldnt have the shutdown temp set by the old standard would they now ? (if you read the thread you will see that many people had serious problems with comps just shutting down while they didnt even overclock)

but even if, Im currently on page 7 of that thread and Im gonna quote someone from there:

"..... i would take the lower (not really what core is temp) but consistant reading across the board that can be compared with other setups to judge the effiency of my system over the random who knows what it is reading we are getting. since the data will match up with a plethera of others systems i will still know what the failure temp is but i will be better able to troubleshoot and judge my cooling systems effectiveness cause im not the pink flamingo in a sea of white flamingos(meaning abit only one with this crazy problem)"

page 6, posted by "maximus x-d"



D

 

[Peter]

Temperature readings aren't made to be compared with other setups. As I said already, the measurement methods differ too wildly to be anywhere near comparable.

Abit's point here being: They read the temperature at the exact spot AMD uses for their definition of maximum operating temperature. This means the AMD defined shutoff temperature threshold can be used.

Yes we're used to something else. Something much less useful and accurate. Something that gives too comforting a reading. Something that is way slower in reacting to overheating conditions. Get over it.

Whether or not Abit got the thresholds thing right is a different question.

 

[ZL1]

Originally posted by: Peter
Temperature readings aren't made to be compared with other setups. As I said already, the measurement methods differ too wildly to be anywhere near comparable.

Abit's point here being: They read the temperature at the exact spot AMD uses for their definition of maximum operating temperature. This means the AMD defined shutoff temperature threshold can be used.

Yes we're used to something else. Something much less useful and accurate. Something that gives too comforting a reading. Something that is way slower in reacting to overheating conditions. Get over it.

Whether or not Abit got the thresholds thing right is a different question.

ok I got the whole thing
see the fact that at this time direct die reading is not the standard reading method confused me the most (after all Intel's own chart is in IHS surface temps and this same chart is the point of reference for most people) + the deal with the shutdowns didnt help either (this one supported the whole "wrong readings" theory)


D