cancel
Showing results for 
Search instead for 
Did you mean: 

Graphics Cards

Swchepple
Journeyman III

New PC - Low GPU utilization+fps issues - 6900 XT defective? - How to troubleshoot?

Description of Original Problem: Noticed that fps would tank drastically after some daily use on pc. In certain games where I normally expect high fps would be cut down by more than half during this my issue. GPU utilization would drop from 80-90%'s to 50-60%'s with obvious fps decrease. Some games are still playable while others become unplayable due to stutters and severely low fps (165fps->60fps). Restarting the PC would generally resolve the issue until it occurs again.

Timing is intermittent, haven't been able to reliably reproduce/trigger issue on demand.

See example in borderlands 3, where in the same scene of benchmark I would receive 250~fps compared to 160~fps when issue occurs. GPU utilization when 250fps is around 80-90's while 160fps is closer to 40-60's.

250~FPS Borderlands 3 Benchmark

160~FPS Borderlands 3 Benchmark

Using adrenalin stress testing, I have image of when GPU is working as expected and when it's a "problem". I noticed that during the "problem" stage, clock speeds and power consumption flutuates and spikes sporadically. See images for reference.

Stable clock speed = high fps in games 

Fluctuating and spiking clock speeds = low gpu util and low fps in games 

I have monitored CPU during these times and while CPU usage is around 40%, none of the cores reach 100%, so I think it is not a cpu bound situation that is causing lower usage in GPU. I could be wrong, I am not exactly super tech savvy.

My question is "How should I go about continuing to troubleshooting this issue? & How can I determine if there is something wrong on the hardware side AKA defective?"

Troubleshooting: Things I have tried:
-Set minimum clock to 100MHz below the max clock speed in adrenalin (clock speeds go up but usage still remain low)
-Disable Deep Sleep States using MPT
-DDU in safemode and reinstall chipset and GPU drivers
-Tried adrenalin versions 22.5.1 and 22.11.2 (Issue occurs in both versions)
-Using different power plans on PC between balanced and high-performance
-Disable Ultra Low Power State (ULPS) in registry
-Disable TPM and fTPM in bios
-Turning off CPU overclock and running in ECO mode
-Resetting PC using Windows
-Reinstalling Windows via Media Creation Tool USB

Computer Type: Desktop
GPU: ASUS TUF GAMING OC Radeon RX 6900 XT 16 GB
CPU: RYZEN 5 7600x 6 CORE 12 THREADS
Motherboard: ASUS TUF x670E-PLUS WIFI
BIOS Version: 0821
RAM: 32GB G.Skill Trident Z5 Neo RGB 6000MHZ CL36 32GB
PSU: Corsair RM1000x 1000W 80+ Gold Fully Modular
Case: Lian Li O11 Dynamic Evo
Operating System & Version: WINDOWS 10 PRO 22H2 19045.2364
GPU Drivers: AMD Adrenalin 22.5.1 (WHQL)
Chipset Drivers: AMD x670E CHIPSET DRIVERS VERSION 4.11.15.342
Background Applications: DISCORD, CHROME, STEAM

0 Likes
3 Replies
Swchepple
Journeyman III

Is this hardware defect or can I do something software side to circumvent this? 

0 Likes
xlox
Forerunner

did you check temps while gaming with adrenalin superposition ? if junction temps gets to 110 it could be throttling ...

but the "issue" can be coming by different causes, windows 10 can cause this depending on games (had problems with gears5 because of an obscure setting (wintimer hpet ..activated by asus ai) of windows ), monitor sync can cause this etc

0 Likes

junction/hotspot temps hover 75C when actually in use and CPU doesn't exceed 60C. When in downclock state, it goes to 40C and stays there, fans aren't in use unless I tune it to do so (obv not needed).

 

0 Likes