search

harddiskdrive / open-hardware-monitor

Automatically exported from code.google.com/p/open-hardware-monitor
0 stars 1 forks source link

Incorrect readings on ASUS M4A79XTD EVO motherboard #79

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
What is the expected output? What do you see instead?

I am using Open Hardware Monitor along with Asus PC Probe to monitor my system, 
and they give different results.

The temperatures that PC Probe reports as CPU and MB are being reported by OHM 
as Motherboard temperature 1 and 2.  The temp reported by OHM for CPU 
tremperature is very different (by 11 degrees C) from the one being reported by 
PC Probe. 

I have no way of knowing which is correct, but since ASUS built the board and 
also supplied the PC Probe software, I'm thinking that they are likely to be 
correct.

I've attached a screen capture showing the two sets of readings

---

What version of the product are you using? On what operating system?

v0.1.35 beta, Windows 7 Home Premium 64 bit

---

Please provide any additional information below.

The motherboard model is correctly identified by OHM

---

Please attach a Report created with "File / Save Report...".

attached

Original issue reported on code.google.com by carlson....@yahoo.com on 25 Jun 2010 at 12:43

Attachments:

GoogleCodeExporter commented 9 years ago 
Thank you.

The temperatures measured by the IT8720F chip need not to be all from the 
sensors on the mainboard. One of the temperatures is usually a sensor in (or 
near) the CPU. 

So "Temperature #1" is "CPU" and "Temperature #2" is "Mainboard" in your case. 
The core temperatur readings from the CPU are not shown in Asus PC Probe. The 
core sensors react very fast to temperature changes in the CPU core, but they 
might not be very accurate when you want absolute values. So you might need to 
add a small offset.

Does Asus PC Probe show any other sensors like for voltages or fans you haven't 
connected? If you can provide a screenshot of all sensors of Asus PC Probe and 
an OpenHardwareMonitor.Report.txt saved approximately at the same time, then I 
can add special code to correctly label all the sensors from the IT8720F chip 
on your mainboard.

Original comment by moel.mich on 25 Jun 2010 at 6:48

GoogleCodeExporter commented 9 years ago 
Thanks.  Here are the items you asked for.  I turned on all of Probe's readings 
and captured it alongside OHM and also saved a report.

Original comment by carlson....@yahoo.com on 27 Jun 2010 at 3:23

Attachments:

GoogleCodeExporter commented 9 years ago 
Thank you. I have tried to identify the sensor channels of the IT8720F chip. I 
am still a bit unsure if the CPU VCore voltage and the +3.3V voltage are 
correct. Is the VCore really 0.97V, does it change depending on load?

In case you can record another screenshot and report with slightly changed CPU 
VCore that would be helpful. But be careful, only change the CPU VCore if you 
know what you are doing, otherwise you can easily damage your CPU.

Here is the new version to test: 
http://openhardwaremonitor.org/openhardwaremonitor-v0.1.36-alpha.zip

Original comment by moel.mich on 27 Jun 2010 at 8:47

GoogleCodeExporter commented 9 years ago 
I'm reluctant to modify VCore myself, but it does change with load -- AMD's 
"Cool and Quiet" feature does that.  The 0.97 reading was probably taken with 
the system idle, so I ran another test with the new 1.36 alpha version and with 
the system working a bit.  I'm attaching the results.  It looks like OHM's 
reading of VCore never varies from 0.98, even when PC Probe says otherwise.

Original comment by carlson....@yahoo.com on 29 Jun 2010 at 12:49

Attachments:

GoogleCodeExporter commented 9 years ago 
Thank you very much. So it seems that the CPU VCore, +3.3V, and +12V are not 
monitored with the IT8720F chip, but with the AMD southbridge (cf. attached 
CPU-Z report from another system) for which there is no support in the Open 
Hardware Monitor.

I have hidden the wrong values now, which completes the mainboard specific 
configuration for the IT8720F on the ASUS M4A79XTD EVO.

You will find the new version here: 
http://openhardwaremonitor.org/openhardwaremonitor-v0.1.36-alpha.zip

Original comment by moel.mich on 29 Jun 2010 at 7:54

Attachments:

GoogleCodeExporter commented 9 years ago 
Something very strange has suddenly started happening in my system.  I had 
installed OHM 1.36 and later I rebooted and since then I have been getting 
bizarre and conflicting readings from both OHM and PCProbe.  I've attached and 
example showing OHM 1.35, OHM 1.36 (the first one you sent, not the latest -- I 
haven't had a chance to try the latest) and PC Probe.  The results, especially 
the temperatures reported by Probe and the +5v reading from both, were so 
alarming that I shut the system down to make sure I wasn't damaging anything.

Can you explain this?  Could OHM have done anything to change the settings on 
the motherboard, or to change the values that PC Probe is reading?  How can I 
fix whatever is going on here?

Original comment by carlson....@yahoo.com on 29 Jun 2010 at 12:21

Attachments:

GoogleCodeExporter commented 9 years ago 
This is nothing to worry about, it is caused by the fact that the access to the 
IT8720F chip is not synchronized. So the Open Hardware Monitor might tell the 
IT8720F that it wants to read voltage #1, then comes Asus Pc Probe and tells 
the IT8720F it wants to read temperature #1, then the Open Hardware Monitor 
reads the value from the IT8720F thinking it is a voltage, but it actually gets 
the value for the temperature. And of course the other way as well resulting in 
wrong values in Asusus PC Probe. The result is completely wrong values in all 
monitor applications involved.

Solving the problem is easy. Just run only one monitoring application at a time.

If you start the application by hand very often you are lucky and the time 
windows where they access the IT8720F do not overlap. But of course if you 
autostart the applications after boot they get started almost perfectly at the 
same time, so there is a big chance that their time windows overlap. And of 
course if you run three monitoring applications chances increase once more that 
one disturbs the other.

Original comment by moel.mich on 29 Jun 2010 at 1:07

GoogleCodeExporter commented 9 years ago 
I understand that explanation and I hate to keep pestering you with questions, 
but I think I continued to see the strange readings in PC Probe even after I 
closed OHM.  Does that make sense?  (I'm not at the system now so I can't 
re-check, but I'm pretty sure it's true.)

(My concern here is to be sure that the crazy temperatures and voltages are 
truly a mis-reading and nothing has gotten broken in such a way that the 
setting have changed and the system is really running at damaging voltages and 
temperatures.  I appreciate your patience in answering my questions.)

Original comment by carlson....@yahoo.com on 29 Jun 2010 at 1:46

GoogleCodeExporter commented 9 years ago 
If you really closed all monitoring applications (this includes others as well 
of course) except one, then the remaining one should display correct values 
once more. Thats at least the case if you run multiple instances of the Open 
Hardware Monitor only. 

I am not sure what Asus PC Probe does internally, so I can't give any real 
answer for that case. It could happen that the hardware monitoring chip gets 
somehow misconfigured, but in that case a reboot should fix the problem 
(because the settings are not persistent). 

In any way, I am sure that your system is not running at 128°C or 6.85V on the 
+5V line. You get these values if you read the unused voltage channels (with 
values 0xFF = 255) and convert them to temperature or +5V voltage signals. Your 
screenshot also nicely confirms that the CPU VCore, +3.3V and 12V signals are 
read from a different chip, because there the values are still correct (because 
no collisions happen there).

Original comment by moel.mich on 29 Jun 2010 at 2:16

GoogleCodeExporter commented 9 years ago 
I downloaded and ran the latest 1.36 version and it looks good.  For whatever 
reason, it didn't seem to interfere with PC Probe, but I'll take your advice 
and run only one or the other from now on.  When I had both of them running, 
the readings seemed to match in all cases.

Thanks again for your help and support.

Original comment by carlson....@yahoo.com on 30 Jun 2010 at 1:23