GPU Usage Over 100%, low fps in game

[WildcatsMike]

This is a problem I started having a couple weeks ago, and it's one I haven't figured out how to solve.

So I booted up GTA V, let it load, and it played fine at 60 fps. I played it for a little bit, and all of the sudden, the framerate nearly halves itself, reaching a max of 40, and an average of about 30ish. When rebooting the game, it still stays at 30ish.

I looked into the issue, and every solution I found failed to solve the issue. I did some monitoring, and the GPU usage after the drop often got to over 100%. Gtav.exe was using about 90%, but oddly, gtavlauncher.exe kept popping up with around 15% usage, and it kept pushing the usage over 100%. It doesn't constantly stay over 100%, but the framerate still suffers constantly in ALL of my games. The CPU and Memory usage never max out when the GPU usage is over 100%.

A couple of interesting things to note:

-MSI Afterburner reports that my gpu usage when idle is ~95%+. I've used Process Explorer to check this, and it reports <1% usage when idle. Probably just a bug with Afterburner, but just thought I'd add that in there.

-These problems started just after I reseated my CPU cooler, as the temps were reaching a bit too high (Maxed @ 72°C, at a constant ~60°C). I've since reseated it again, as I was unhappy with the first reseating. Temps were improved.

-When reseating the CPU the 2nd time, I upgraded from 8gb to 16gb of RAM, and installed a better 120mm case fan.

PC Specs: http://pcpartpicker.com/list/KJwkLD

Any help is greatly appreciated.

 

[maxwellmelon]

You fail to say what your GPU temps are during gaming. Playing fine for a little bit then droping is a big sign of thermal throttling.

 

[WildcatsMike]

You fail to say what your GPU temps are during gaming. Playing fine for a little bit then droping is a big sign of thermal throttling.

GPU temps go no higher than 63°, averaging 50-60° when running fine. When running poorly, GPU temps are between 64-66° constantly.

 

[maxwellmelon]

ok just to check (and you might have done it this way) you say you fixed your cpu temp was overheating but did you check the thermal margin??
The proper way to read the AMD FX cpu die temp is though amd overdrive and use its "thermal margin" is the thermal margin a positive number not close to 0. and the bigger the number the cooler the cpu die is. if its a negative number then your cpu die is considered to be overheating. chances are based off of how you wrote that number that is actually the socket tempeture (the cpu die will be hotter then that) (I have an fx8350 and I love amds fyi just not sure why they went down this path for temp speed control)

more info http://www.tomshardware.com/faq/id-2122665/understanding-temperature-amd-cpus-apus.html

 

[WildcatsMike]

ok just to check (and you might have done it this way) you say you fixed your cpu temp was overheating but did you check the thermal margin??
The proper way to read the AMD FX cpu die temp is though amd overdrive and use its "thermal margin" is the thermal margin a positive number not close to 0. and the bigger the number the cooler the cpu die is. if its a negative number then your cpu die is considered to be overheating. chances are based off of how you wrote that number that is actually the socket tempeture (the cpu die will be hotter then that) (I have an fx8350 and I love amds fyi just not sure why they went down this path for temp speed control)

more info http://www.tomshardware.com/faq/id-2122665/understanding-temperature-amd-cpus-apus.html

Thermal Margin ranges between 8-16°C, never goes negative. This is while things run poorly. Haven't been able to test it when things run fine, as I can't seem to get things to run fine at all anymore.

 

[maxwellmelon]

ok just to check (and you might have done it this way) you say you fixed your cpu temp was overheating but did you check the thermal margin??
The proper way to read the AMD FX cpu die temp is though amd overdrive and use its "thermal margin" is the thermal margin a positive number not close to 0. and the bigger the number the cooler the cpu die is. if its a negative number then your cpu die is considered to be overheating. chances are based off of how you wrote that number that is actually the socket tempeture (the cpu die will be hotter then that) (I have an fx8350 and I love amds fyi just not sure why they went down this path for temp speed control)

more info http://www.tomshardware.com/faq/id-2122665/understanding-temperature-amd-cpus-apus.html

Thermal Margin ranges between 8-16°C, never goes negative. This is while things run poorly. Haven't been able to test it when things run fine, as I can't seem to get things to run fine at all anymore.

understand that. but you now verified for sure that your issue is not caused by thermal throttling of the cpu or gpu.

next thing to try would be to uninstall and then reinstall your video drivers, or try a different driver version. other then that idk as these are the most common issues.

 

[WildcatsMike]

ok just to check (and you might have done it this way) you say you fixed your cpu temp was overheating but did you check the thermal margin??
The proper way to read the AMD FX cpu die temp is though amd overdrive and use its "thermal margin" is the thermal margin a positive number not close to 0. and the bigger the number the cooler the cpu die is. if its a negative number then your cpu die is considered to be overheating. chances are based off of how you wrote that number that is actually the socket tempeture (the cpu die will be hotter then that) (I have an fx8350 and I love amds fyi just not sure why they went down this path for temp speed control)

more info http://www.tomshardware.com/faq/id-2122665/understanding-temperature-amd-cpus-apus.html

Thermal Margin ranges between 8-16°C, never goes negative. This is while things run poorly. Haven't been able to test it when things run fine, as I can't seem to get things to run fine at all anymore.

understand that. but you now verified for sure that your issue is not caused by thermal throttling of the cpu or gpu.

next thing to try would be to uninstall and then reinstall your video drivers, or try a different driver version. other then that idk as these are the most common issues.

This seems to have fixed the problem after 3 different boots and putting the PC under load for a while. If the issue comes back, i'll edit my post. If the problem doesn't come back, thanks for the help! 😀

 

[DexterMullen]

I am on desktop and I got rtx 2070, got the same problem, solved it by changing in nvidia control panel /manage 3d settings/power management to performance.
It looks like when some loading in games appears and gpu usage goes to very low, but then it stays in that mode after loading is done.
But this fixed it for me 100%

http://dodaj.rs/images/fix.jpg