Conflicting temperature readings on FX6300

we r me go boom

Reputable
Jun 7, 2014
5
0
4,510
I have a FX6300 running at stock speeds, with a Corsair Hydro H75, on a MSI 970A-G43 motherboard.

Coretemp says my CPU is idling at 21c
Hardware Monitor says my CPU is idling at 16-24c
AMD Overdrive says my CPU is idling at 49-54c.
And my BIOS read 39c at idle, or I assume is idle, since I just let it sit on it for about 10 minutes.

Everything is, to the best of my knowledge, at stock configuration on my PC, except for the fans, which are all at 100%, with the exception of the CPU coolers fans, which I can't seem to control no matter what I try.

My fan controller says the ambient temp inside my PC is 75F, which correlates to the actual temperature of 75f inside my room.

Which temp reading do I trust?
What's the actual idle temp I should be getting?
Is the temp it's at now, if I believe it to be 49-54c, safe for it?
 
Solution
@damric: Here is my edit for your tutorial.


Does HWMonitor say your AMD CPU/APU temperature is 60 °C but the CPU is throttling performance already? Is Core Temp only reading 10 °C but it's not possible because your room temperature is 28 °C? The reason is that the temperature values shown by AMD CPUs are not in °C. They are on a different temperature scale made by AMD. That means a temperature value of 30 should be considered as 30 AMD temperature units or AMDT units in short and not in degrees Celsius. Values shown by Intel CPUs are in actual °C.

The proof for this statement can be taken from one of AMD's technical documents. The technical document is meant to be read by engineers so we will highlight the part that matters to...

we r me go boom

Reputable
Jun 7, 2014
5
0
4,510
I guess I'll run Minecraft and blow up a bunch of TNT to put stress on it, then, since that's the most CPU intensive thing I have right now, I think.

Which, probably isn't very CPU intensive at all, I guess, since Core Temp didn't really change when while I blew up 40 TNT.

During the explosion, core temp went to 27c.


So then I blew up 600 blocks of TNT at once, and Coretemp went to 36c, and now it just seems to be idling at 29c.
 

cball1311

Honorable
Dec 15, 2012
1,622
1
12,160
Here is Prime95 download. Download it, unzip, and run it for about 5 minutes with Core Temp running. Just open it and hit "Ok" and the test should start running. If your CPU temp goes up to 70C stop the test. You can stop the test by going to "Test" menu option, then "Stop...". This will give you true load temps.
 

we r me go boom

Reputable
Jun 7, 2014
5
0
4,510


nah, i just don't think coretemp is a very good program at all, period.

http://i.imgur.com/pNj6Kgn.png

i think i'll trust MSI Control Center over it, since its temp reading actually sounds plausible.
 


Although I up-voted your previous post and the tutorial also tells the correct way to monitor AMD CPU temperatures, I don't like that tutorial because it is very AMD biased.

Here are 2 sentences from the tutorial:

"Well if these 3rd party programs worked like they say they are supposed to work, we wouldn't be having this problem."

"This is also the signal that is most often erroneously interpretted by 3rd party programs like HWMonitor, ect."

Both these statements are in favor of AMD. Third party programs are doing their best they can to show accurate readings based on what AMD is giving them. Those same programs show almost correct readings for Intel CPUs. Also you won't find temperature benchmarks for AMD CPUs but there is for Intel CPUs. Ever questioned why AMD Overdrive now shows Thermal Margin instead of CPU temperature? So that you will never know the current temperature that AMD never gets it right
 


Yeah, I know, I just had to repost it since it seems that the OP completely neglected it. The tutorial still shows the only accurate way to measure the temps (well, the thermal margin at least), and it also explains why the readings can be so conflicting when looking at different programs. It's not the tutorial's fault that AMD has to make measuring the temps so hard.
 
No I meant the tutorial's author is blaming the third party programs for not showing accurate readings when it's not their fault.

For example,
this is from CoreTemp:
http://www.alcpu.com/CoreTemp/howitworks.html

"Intel defines a certain Tjunction temperature for the processor. This value is usually in the range between 85°C and 105°C. In the later generation of processors, starting with Nehalem, the exact Tjunction Max value is available for software to read in an MSR (short for Model Specific Register).
A different MSR contains the temperature data. The data is represented as a Delta in °C between current temperature and Tjunction.

So the actual temperature is calculated like this 'Core Temp = Tjunction - Delta'

The size of the data field is 7 bits. This means a Delta of 0 - 127°C can be reported in theory. In fact the reported temperature can rarely go below 0°C and in some cases (Core 2 - 45nm series) temperatures below 30° or even 40°C are not reported."



"AMD processors report the temperature via a special register in the CPU's northbridge. Core Temp reads the value from the register and uses a formula provided by AMD to calculate the current temperature.
The formula for the Athlon 64 series, early Opterons and Semprons (K8 architecture) is: 'Core Temp = Value - 49'.
For the newer generation of AMD processors like Phenom, Phenom II, newer Athlons, Semprons and Opterons (K10 architecture and up), and their derivatives, there is a different formula: 'CPU Temp* = Value / 8'.

*CPU Temp is because the Phenom\Opteron (K10) have only one sensor per package, meaning there is only one reading per processor."

As you can see reading values for Intel CPUs are way much logical than AMD's
 


The vast majority of AMD users are not aware of this, most just trust the temp readings given by the software, so I doubt AMD will do anything about it.
 


Well in the past they have fixed one thing

[flash=560,315]https://www.youtube.com/v/BSGcnRanYMM[/flash]

Here they fixed it, thanks to tomsHardware

[flash=560,315]https://www.youtube.com//v/cAqlA9EJ4ME[/flash]

 
That tutuorial is only meant for AMD CPU/APU. I do need feedback to revise the FAQ to make it better. That rev 1 straight off the top of my head when I wrote that, (after answering the question for the 12th time that day), so yeah it could probably use some editing :D

Anyways, is the OP's temp figured out now? It looks like it to me...
 

dacquesta1

Honorable
Jan 1, 2014
394
0
10,810
OP...run p95 for more than a few seconds on the first run and see what your temps go up to. If you want to know your max load temps you need to let prime run for at least 15-20 minutes. MSI program is probably reading what what hwmonitor shows as CPU temp and coretemp shows what hwmonitor shows as package temp.
 
@damric: Here is my edit for your tutorial.


Does HWMonitor say your AMD CPU/APU temperature is 60 °C but the CPU is throttling performance already? Is Core Temp only reading 10 °C but it's not possible because your room temperature is 28 °C? The reason is that the temperature values shown by AMD CPUs are not in °C. They are on a different temperature scale made by AMD. That means a temperature value of 30 should be considered as 30 AMD temperature units or AMDT units in short and not in degrees Celsius. Values shown by Intel CPUs are in actual °C.

The proof for this statement can be taken from one of AMD's technical documents. The technical document is meant to be read by engineers so we will highlight the part that matters to us.
"Tctl is a processor temperature control value used for processor thermal management. Tctl is accessible through D18F3xA4[CurTmp]. Tctl is a temperature on its own scale aligned to the processors cooling requirements. Therefore Tctl does not represent a temperature which could be measured on the die or the case of the processor. Instead, it specifies the processor temperature relative to the maximum operating temperature, Tctl,max."

900x900px-LL-306a7aae_tclscale.png

Now the problem is a value shown in AMDT units is of no help to us because it doesn't tell the real temperature. One thing we do know that this value needs to be less than the maximum temperature limit for your CPU defined by AMD. For example, here are some maximum temperature limits(shutdown temperatures) for some AMD CPUs:

FX-8350 : 70 AMDT units
FX-6300 : 90 AMDT units

The less the value from the maximum temperature limit, the better. Now there is another problem. All information related to temperatures such as maximum temperature limits are nowhere to be seen on AMD's official website, not even in technical documents. You can't just google and trust some random source for maximum temperature limit for your CPU. The third party program CoreTemp shows maximum temperature limits as Tjmax for all CPUs.

AMD's own offering - AMD OverDrive now shows Thermal Margin for the CPU. "Starting with AMD OverDrive Ver 4.3.1.0690 instead of displaying CPU temperature, AOD will report Thermal Margin. Thermal Margin indicates how far the current operating temperature is below the maximum operating temperature of the processor." So basically a higher value of Thermal Margin is better. AMD still shows °C for Thermal Margin but it should still be considered in AMDT units. Thermal Margin is probably the best way to monitor your AMD CPU temperatures which already takes into account the max temperature limit and shows the difference.

Download AMD Overdrive

[rest as is]
 
Solution