[SOLVED] Random stuttering in games, tried almost everything, not sure where to go next.

Sep 5, 2020
1
0
10
SPECS:
  • Case: Cooler Master HAF-XB Evo
  • Motherboard: Asus Prime X570-P
  • CPU: AMD Ryzen 7 3700X w/ Wraith Prism Cooler
  • RAM 2x16GB G. Skill Trident-Z @ 3200Mhz
  • GPU: EVGA RTX2060 KO
  • PSU: Corsair TX750M
  • STORAGE: 1xM.2 NVME, 1xSATA SSD, 1xHDD
  • OS: Windows 10
Bit of a long post, but this has been a journey and I'm not quite sure where the solution to my problem lies, so I figured i'd seek some assistance here.

As the title states, I've recently noticed random and occasional stuttering while playing games, It never really bothered me until I started playing VR games like beat saber where computer stuttering often means missing enough notes to fail the song. What will happen is the game will run fine for 5-10 minutes or so, and then it will stutter for 3-5 seconds before returning to normal. This issue isn't isolated to Beat Saber / VR Games in general, as I've noticed it in other Non-VR games like Satisfactory. However the issue is not present in less-intensive games. I was almost immediately able to eliminate "not meeting game requirements" because the issue is completely random in when it shows up, and isn't tied to sections of gameplay that have lots of lighting/particle effects/other things that would bog a GPU down (plus, with the specs I listed below, I should have absolutely no problem running beat saber).

I delved a bit deeper, realizing that this could be a much more technical issue than I previously thought. I downloaded HWMonitor and started paying close attention to it while games were running. Here are some things I noticed:
  • GPU Usage drops significantly when this happens
  • CPU clock drops with it.
  • Power drops to both CPU and GPU when this happens (Not sure if this is a cause, or an effect)
  • Temperatures are stable throughout the whole process (<75C)
I tested the usual suspects when it comes to shoddy performance, tweaking bios settings like DOCP, things like that, but to no avail.

I started reading threads on similar issues and it seemed like the logical conclusion was that the motherboard VRMs were causing CPU throttling, which led to GPU stuttering. So I thought maybe my case didn't have proper airflow. I added an extra 120mm fan as an exhaust (I didn't have any extra fans aside from what came stock with the case prior to adding this one), but that didn't seem to have much of an affect on thermals. I next tried opening up the case (removing all the panels on the case I use turns it into a test bench) and having a box fan blow air over the components while playing. This also had a minimal affect. What really confuses me is that none of my motherboard temperature sensors report anything above the 70 degree area, which should be well within optimal operating temperatures for the hardware I have. I've modified fan curves to make sure its not GPU/CPU based throttling, and their temperatures are stable and within optimal ranges, so needless to say I'm at a loss. If anyone has any idea what could be going on, I'd be very appreciative of some direction.


Things I've tried (in no particular order):
  • Running games at different settings
  • Updating / Rolling back drivers
  • Clean OS install
  • Actively cooling VRMs
  • Enabling / Disabling D.O.C.P.
  • Different case fan configurations
  • Storing games on different drives
  • Tuning fan curves
  • Plugging PC into different surge protector / wall outlets.
Other Thoughts / Things I haven't tried:
  • Maybe the motherboard I have doesn't have a dedicated VRM temperature sensor, so it never showed up in testing (They're not clearly labeled, mostly just TEMPIN{sensor_number} )
    • If this were the case, I would think my additional cooling attempts would have shown some difference, but no.
  • Maybe the system is drawing above what the PSU can provide, but I would think 750W is enough to power the system with room for overhead, and I tested the 3 rails to make sure the PSU is healthy.
    • Looking at HWMonitor, the sum of max power draw across all components is less than 750W, so I don't know how we could be drawing above the PSU limit.
I can provide the graphs from HWMonitor if that would be beneficial, maybe you'll catch something I missed?

TL;DR: Game stutters every couple of minutes while playing intensive games, not sure whats causing it, tried many things, to no avail.
 
Solution
litterally just made an account to suggest something since i also had this issue long ago,
Open Device Manager​
Go To IDE ATA/ATAPI controllers​
Disable any with name AMD SATA Controller, i personally have 2 copies of this listed.​

thought i'd better explain how i found it.
my keyboard kept disconnecting and reconnecting and this caused the stutter
tried a different keyboard and it stuttered when connected it
went into device manager and noticed another stutter and realised when device manager was populating it would consistantly stutter so it had to be something in there
with a bit of searching around for each device and a bit of trial and error realised disabling those particular devices completely stopped the issue.
Don't know if this suggestion will help, but have you tried another, known, good, working Graphics card? This sounds like un-montiored part of the graphics card might be getting too hot.
And another suggestion, though they might not have any input, talk to EVGA support; they have helped me several times in the past.
 
Sep 8, 2020
1
0
10
litterally just made an account to suggest something since i also had this issue long ago,
Open Device Manager​
Go To IDE ATA/ATAPI controllers​
Disable any with name AMD SATA Controller, i personally have 2 copies of this listed.​

thought i'd better explain how i found it.
my keyboard kept disconnecting and reconnecting and this caused the stutter
tried a different keyboard and it stuttered when connected it
went into device manager and noticed another stutter and realised when device manager was populating it would consistantly stutter so it had to be something in there
with a bit of searching around for each device and a bit of trial and error realised disabling those particular devices completely stopped the issue.
 
Last edited:
Solution