• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

vega 64 gpu frequency problem

chumlee

New Member
Joined
May 7, 2024
Messages
28 (0.07/day)
Hello everyone ,
I noticed since yesterday that my GPU no longer goes up to this maximum frequency which is 1630 MHz and which remains stuck between 800 MHz and 1400 MHz without being stable. However my card is used at 99% during the tests.
I have not changed the component, I have had the same configuration since 2019, however it was going up to 1630mhz before. I also noticed even if I increase the consumption to + 50% my card does not exceed 200w whether it is without oc or with oc. I did several tests on furmark, 3D bench and it remains stuck at the frequency that I mentioned above

I tried several times to run ddu to properly reinstall my drivers but nothing changed.
I checked on Windows that there is no limit or energy saver.
I also tried putting the p7 in the main frequency at minimum and maximum but always the same result. Except when the bench is not launched my frequency is at the maximum.
And all this whether on adrenaline or overdrivetool.
my drivers are up to date, same for the bios which is in f64 version

My current configuration:
Motherboard: aorus 470x ultra gaming
processor: r7 2700 oc at 3.9GHz (base 3.2GHz)
gpu: asus rog vega 64 strix oc 8 gb
ram: DDR4 HyperX Predator RGB, 16 GB (2 x 8 GB), 2933 MHz, CAS 15
psu: EVGA SuperNOVA 750 G3, 750W
processor cooling: iCUE H100i ELITE


Thank you for your help, hoping to find a solution to my problem.
 

Attachments

  • gpuz.JPG
    gpuz.JPG
    62.6 KB · Views: 70
  • furmark bench without oc or  power limit changed.JPG
    furmark bench without oc or power limit changed.JPG
    437.2 KB · Views: 72
  • bench with UV.OC power limit +50.JPG
    bench with UV.OC power limit +50.JPG
    490.4 KB · Views: 70
  • only p7  without benchmark.JPG
    only p7 without benchmark.JPG
    183.9 KB · Views: 73
  • only p7 with benchmark.JPG
    only p7 with benchmark.JPG
    376.1 KB · Views: 73
Hello everyone ,
I noticed since yesterday that my GPU no longer goes up to this maximum frequency which is 1630 MHz and which remains stuck between 800 MHz and 1400 MHz without being stable. However my card is used at 99% during the tests.
I have not changed the component, I have had the same configuration since 2019, however it was going up to 1630mhz before. I also noticed even if I increase the consumption to + 50% my card does not exceed 200w whether it is without oc or with oc. I did several tests on furmark, 3D bench and it remains stuck at the frequency that I mentioned above

I tried several times to run ddu to properly reinstall my drivers but nothing changed.
I checked on Windows that there is no limit or energy saver.
I also tried putting the p7 in the main frequency at minimum and maximum but always the same result. Except when the bench is not launched my frequency is at the maximum.
And all this whether on adrenaline or overdrivetool.
my drivers are up to date, same for the bios which is in f64 version

My current configuration:
Motherboard: aorus 470x ultra gaming
processor: r7 2700 oc at 3.9GHz (base 3.2GHz)
gpu: asus rog vega 64 strix oc 8 gb
ram: DDR4 HyperX Predator RGB, 16 GB (2 x 8 GB), 2933 MHz, CAS 15
psu: EVGA SuperNOVA 750 G3, 750W
processor cooling: iCUE H100i ELITE


Thank you for your help, hoping to find a solution to my problem.
It could be a protection from the driver to prevent damaging the card when using furmark. Did you try a normal game?
 
I launched red dead redemption 2 to check and I'm running around 1000mhz - 1400mhz I noticed this problem yesterday with vrchat in vr mode where I was running between 800 and 1000 mhz

I also used 3Dbench and heaven and I had the same result
 

Attachments

  • RDR2.JPG
    RDR2.JPG
    407.9 KB · Views: 54
It's probably the Hotspot temperature, 105C is too hot so it's throttling the GPU to keep the temps from rising further. Normal Hotspot temp for Vega is about 10-20C above GPU temp.

Time to change the thermal paste.
 
you have a hotpsot delta of 40°C.
Repaste your GPU (be careful! the interposer breaks very easily if someone with zero experience does it)
 
you have a hotpsot delta of 40°C.
Repaste your GPU (be careful! the interposer breaks very easily if someone with zero experience does it)
Thank you for your answer, I already had it at the time, by changing the thermal pad and the thermal paste but unfortunately from what I read the vega 64 heats up enormously and it didn't change much
 
Are the pads all exactly the right thickness?
My Vega 64 has a ~15°C delta.
There is something wrong.
 
Hi! I come back to you as planned, so I changed the thermal pad I noticed that on the old pad I had taken 2mm or even 2.5mm and this time I took 3mm and I also changed the thermal paste.
Unfortunately I was not able to remove the plate on the PCB to change the two small pieces of thermal pad unfortunately I only managed to remove 3 screws on the whole and I preferred not to force it so as not to damage it.

So I did two tests one without UV and another with UV and I noticed a drop on the hotspot the pick without UV is 90-92
And with UV of 75-80 and use of power draw at 240w. and all this with furmark.
The hotspot without benching in normal use is 27°C
Do you think the hotspot in both cases is good or is this one still too hot?

I will certainly provide feedback regarding in-game usage to see if there is a problem or not.

Thank you for your feedback.

Sincerely.
 
Last edited:
still pretty hot but as long as the card works as intended (does it?) it's acceptable.
 
Back
Top