GPU sensors going blank/gray

Not exactly sure when this started happening, but the GPU clock information will occcasionally go blank for a split second on the main window, and all clock information goes gray on the sensor window. When the values return, the "GPU" text remains gray. Almost looks like the GPU is disabled afterwards.

Before values go blank
VLoSvpB.png

After
FcB9sp5.png

Sensor Window

frbovlp.png


I know this seems like a pretty harmless issue to point out, but the same thing happened to the VRM sensor reading on my previous motherboard that died a few days later. Just wondering if any of this could be related to a hardware issue, because I've been getting some odd nvlddmkm driver crashes recently. The card also randomly throttled itself to idle clocks for the first time yesterday even though I was within thermal limits.

Here are my PC specs:
zvGT3Cw.png

PSU is a Corsair RM850x.
 
Grey values in sensors window mean HWiNFO is unable to read a valid value from the sensor. In case of GPUs this usually means it's in a "sleeping" state with clock gated.
 
Grey values in sensors window mean HWiNFO is unable to read a valid value from the sensor. In case of GPUs this usually means it's in a "sleeping" state with clock gated.
I see. It seems to happen at completley random times though. Is it normal for a GPU to enter this sleep state while under load?
 
Under load it's not normal, so it might rather be an issue in GPU drivers. Try to update them.
 
Did that yesterday. Ran DDU and then installed 512.77.
Think I figured out what's causing it. If I have both the "Sensor Status" and "System Summary" windows open, the grey text/blank values issue will occur. If I only have one of the windows open, it doesn't happen. Tested it for a few hours to make sure. Maybe it's a polling problem?
 
Thanks for your observation. Indeed, this could be an issue if both summary and sensors windows are active. Let me have a look at this and see how to fix it.
 
I've observed this same behavior. My hunch is that it has to do with Afterburner running and grabbing GPU clock info at the same time as HWiNFO.

FWIW, it doesn't stay grey.. Most of the time the clock values are in black, but they will consistently go grey.

 
Follow-up: I think I was wrong about my Afterburner theory. Check out this video that I just recorded with only HWiNFO64 running. At the very beginning of the clip you'll see it greyed out, but since the GPU is engaged when encoding the recording it quickly becomes black.

So, is it normal for clock speed to be grey when the GPU is idle?

 
Yes, it might be normal when the GPU is idle that the gfx engine is switched off to preserve power.
 
Hi all (& Martin),
I'm having a similar issue but not w/GPU sensors. For me it's the following sensors, and they only grey out for a few seconds and then come back on:

VCore Latch Min/Max
CPU Core Current
T-sensors (both W_In and T_Sensor headers on my ASUS Maximus Apex XIII) - expressed as Water In T Sensor and Coolant Temp below. Sometimes both grey out, but in the screen shot its just the Water In
CPU OPT header (pump 2 is plugged into this -- a vario pump so its just supposed to read the rpm). Pump 1 is plugged into CPU header, does not grey out
Flow Rate sensor, which is on the W_Flow Header (Koolance sensor)

Weirdly, the Pump 2 Speed (CPU OPT header) will just drop to 0 RPM, and when this happens the flow rate sensor value greys out. The pump does not actually drop to 0 rpm though...

I am not running any other sensor programs aside from Afterburner (and RTSS)

Sensors.PNG
 
I will add that after doing some research, it looks like the sensors I've mentioned are part of the Embedded Controller? I went into HWINFO64 main settings and unchecked the box "EC Support", and a bunch of sensors totally disappeared from the Sensor Status window. Upon closer examination, it was ALL the sensors that I mentioned having an issue with "greying out".

Is there something I can do to rectify this in any settings? Is there a known fix for this? I've also been having some random "hard restart" type crashes with no BSOD lately, during random loads (idle, gaming, starting blankly at my screen) and I read that this might be attributed to EC support. But those seemed like very old threads, so not sure if monitoring EC still has this negative affect.

Any potential solutions would be great! The greying out of the sensors is annoying, but doesn't seem to be a problem (unless of course its related to the system instability, which definitely is a problem and had me hunting for hardware issues - found none).
 
Hi all (& Martin),
I'm having a similar issue but not w/GPU sensors. For me it's the following sensors, and they only grey out for a few seconds and then come back on:

VCore Latch Min/Max
CPU Core Current
T-sensors (both W_In and T_Sensor headers on my ASUS Maximus Apex XIII) - expressed as Water In T Sensor and Coolant Temp below. Sometimes both grey out, but in the screen shot its just the Water In
CPU OPT header (pump 2 is plugged into this -- a vario pump so its just supposed to read the rpm). Pump 1 is plugged into CPU header, does not grey out
Flow Rate sensor, which is on the W_Flow Header (Koolance sensor)

Weirdly, the Pump 2 Speed (CPU OPT header) will just drop to 0 RPM, and when this happens the flow rate sensor value greys out. The pump does not actually drop to 0 rpm though...

I am not running any other sensor programs aside from Afterburner (and RTSS)

View attachment 7927

This issue is usually caused by some other agent/program interfering with HWiNFO when it attempts to read the same sensors. It could be ASUS AI Suite, some other monitoring tool or the mainboard OLED firmware.
 
This issue is usually caused by some other agent/program interfering with HWiNFO when it attempts to read the same sensors. It could be ASUS AI Suite, some other monitoring tool or the mainboard OLED firmware.
Yeah I saw that mentioned in threads here, but I don't have anything else other than Riva running in terms of applications/monitoring, unless there's something I'm not aware of that's reading the same ones. I noticed they all seem to be part of the Asus EC, as disabling EC removes all the same sensor values as the problematic ones. Mainboard OLED I have no idea about, but can look into it.

I will say, for future reference, I uninstalled the regular version of HWINFO64 and instead booted the portable version. Upon loading, I did not see the same issue happening with those sensors, even though moments before with the desktop version it was occurring. Not sure if that's a clue to anything, or if it's just coincidence. I haven't checked on it since earlier today, so I will let you know if I continue to see the improvement using the portable version.

EDIT: I checked again and saw some greying out happening with the portable version too, although seemingly not as frequent and not as many different sensors at a time.
 
Last edited:
There should be no difference in this behavior between the installer and portable version.
I've seen such issues on later AMD mainboards and upon some inspection together with ASUS, we assumed this could be caused by the mainboard firmware accessing the sensors (e.g. to disablay sensor values on the OLED). You might test this by disabling the OLED in BIOS.
 
There should be no difference in this behavior between the installer and portable version.

I thought not, but figured I'd mention it. It was premature anyway, the issue persists.
I've seen such issues on later AMD mainboards and upon some inspection together with ASUS, we assumed this could be caused by the mainboard firmware accessing the sensors (e.g. to disablay sensor values on the OLED). You might test this by disabling the OLED in BIOS.

EDIT: Submitted before I finished typing.

It was happening a lot just recently, so I went into BIOS and this OLED field was set to display "POST codes only". I changed it to "disabled". So far, it hasn't fixed the greying out. It does seem less frequent, but still happening. Mostly just to the temp sensor header values now.
 
Back
Top