Idle Temps on Vishera/Piledrivers?

Darren

Darren

Moderator
Staff member
Since I'm getting an 8320 I've been doing some looking around and noticed that a lot of people have idle temps that seem crazy low, like 16. I've even seen one that was 7. I built 2 friends with computers, each of which using an FX 6300 and they also had REALLY low idle temps, in the teens and low 20's. I wouldn't think that to be possible, let alone on stock cooling. Are the temperature sensors wrong on this line? I did some reading and some mention that the way the temp is calculated for these is messed up at lower load but at higher/full load it's more accurate than previous calculation means.

Any experience or advice with this? More in the wondering then anything else. I know idle temps don't really matter anyway but just curious.

Thanks.
 
raymar said:
nothing wrong with low temps, high is bad. I wouldn't worry about it.
Click to expand...

...

That doesn't answer my question in the slightest. I'm aware that low temps are good. I'm asking WHY do the procs report such low temps. It's curiosity since it's not a physical problem.
 
I haven't messed with AMD chips a whole bunch since my 955 but I remember that most people needed to set an offset in their temp sensor programs for AMD CPUs because they were off. 16c would be 60f. I doubt anyone here keeps their room that cold to maintain that idle temp. I bet Stranghold would know more about if it is off on idle and at load or just off on idle.
 
My 8350 runs at like 10-11 degrees C. Crazy low, curious if that temp is correct.
 
There not correct. I dont know if its the sensor or the bios causing the programs to read it wrong or not. But it does seem to read correct under load. Seems all the bulldozer series are that way. One way thats pretty correct is the motherboard socket sensor reading
 
StrangleHold said:
There not correct. I dont know if its the sensor or the bios causing the programs to read it wrong or not. But it does seem to read correct under load. Seems all the bulldozer series are that way. One way thats pretty correct is the motherboard socket sensor reading
Click to expand...

Okay that's kind of what I was getting from the internet but glad to hear it from you too.
 
http://en.wikipedia.org/wiki/Cool'n'Quiet
read this article, it's the cool and quiet tech that lowers voltage/clock speed when chip is idling (null use) I have a amd quad core on a asus board that drops below 65f, 16c is around 68f I'm sorry I didn't give a better answer before, I thought you just wanted to know if something was wrong.
 
raymar said:
http://en.wikipedia.org/wiki/Cool'n'Quiet
read this article, it's the cool and quiet tech that lowers voltage/clock speed when chip is idling (null use) I have a amd quad core on a asus board that drops below 65f, 16c is around 68f I'm sorry I didn't give a better answer before, I thought you just wanted to know if something was wrong.
Click to expand...

Yeah cool and quiet has been around for a while. I have it on the current chip. Even then there's no way your temps can get that low without some sort of refrigeration. Getting it below room temperature is impossible without something else acting upon it other than air or water cooling.

Cool and quiet would drop this chip when it wasn't overclocked in to the upper 20's. The Piledriver/Vishera just read it wrong. I don't think I've ever seen a chip idle anything under 25 that didn't have a bad sensor.
 
Well I'll just say I see it all the time. Sure the temp sensors aren't 100% accurate, but i'll bet it's close, get a laser temp sensor gun and check it, I have. and overclocking shouldn't apply at idle, unless you've turned off the C&Q. I'm a system builder not a gamer, so I don't get into overclocking anymore except for a local bragging rights contest scrunching seti@home units.
 
Pretty sure that C&Q gets disabled when you start overclocking on your own. At least it does with my board. And thermodynamics wouldn't allow a CPU to get colder than room temperature. It's not possible without some sort of refrigeration or chemical reaction taking place.
 
Has nothing to do with C&Q. One reason, AMD has only one onboard sensor for all cores. So the sensor is not directly on the cores.

(2) No direct reading of physical CPU temperature
The value for the CPU temperature of the (digital) on-chip sensor does not equal any real physical temperature but is measured on an AMD spezific scale. This was done to be able to provide more accurate values when the temperature is close to the thermal design limit of the processor and is used to throttle the CPU to prevent damage by overheating. Therefore reliable measurements for CPU idle states is not possible.

http://help.argusmonitor.com/index.html?TemperaturemeasurementforAMDCPUs.html
 
What's the point in doing that? What's wrong with an accurate measurement? If you don't want people to go to high in temperatures just lower the official safe temperature.

Eh what do I know, just doesn't make sense to do that.
 
Since it is not always apparent whether a particular program is showing the "CPU Core" or "CPU Socket" temperature I compared several programs to find out. Here is what I have concluded:

These show CPU Core Temperature:
HWMonitor: AMD FX-8350 Package Temperature
Angus Monitor: Core 1-8
Core Temp: CPU Temp

These show CPU Socket Temperature
HWMonitor: TMPIN2
Angus Monitor: CPU Socket
EasyTune6: CPU

I ran some tests with an AMD FX8350 (Corsair H60 water cooler):

Idle:
CPU 1409MHz 0.9125V
CPU Core Temperature: 12C
CPU Socket Temperature: 24C

Full Load (using Prime95):
CPU 4206MHz @1.4125V
CPU Core Temperature: 54C
CPU Socket Temperature: 66C
 
You must log in or register to reply here.
Back
Top