Question Random Stuttering on EVGA RTX 3080

Mondeezy

Honorable
Jul 24, 2016
41
1
10,535
0
Hi all,

I just built a new PC over the weekend and it has run smoothly with the exception of random fps stutters when I play video games. It has occurred the most frequently while playing Valorant, but also occurs in other games and I'm not really sure what the cause is. Tried full screen vs windows maximized and tweaking some settings to no avail. Here are my specs:

Specs:
-GPU: EVGA RTX 3080
-CPU: Ryzen 5800x
-Motherboard: MSI MAG B550 Tomahawk
-RAM: 32GB (16x2) of G.Skill DDR4
-Storage: 3TB of mixed storage (SSDs, NVMES)
-PSU: Corsair RM850X
-Display: 144hz ROG Swift PG248Q

UserBenchmark:
https://www.userbenchmark.com/UserRun/47206849

Any help/insight is appreciated. The games pretty much run flawlessly but every couple of minutes will freeze (stutter) for 1-2 seconds before resuming gameplay.

Thanks for your time :)
 

Phaaze88

Titan
Ambassador
1)Cpu.
Chipset driver not installed/updated?
Windows really wasn't carried over from an older drive?

2)980 Pro.
These Gen 4 NVMes need more active cooling than their predecessors did, or else the ASIC controller constantly makes itself thermal throttle.
You'll need hwinfo to see the temperature, as the Magician software doesn't report that sensor. It should appear in hwinfo as Drive Temperature 2.

3)850 Evo.
Perhaps it needs a driver update thru Samsung? Or maybe it's also running toasty?

4)Ram.
Gotta turn on A-XMP first.

Capt. Discombobulate may be right about the latest Nvidia driver. IDK, I'm still on 466.47.
 
Reactions: RodroX and Mondeezy

Mondeezy

Honorable
Jul 24, 2016
41
1
10,535
0
1)Cpu.
Chipset driver not installed/updated?
Windows really wasn't carried over from an older drive?

2)980 Pro.
These Gen 4 NVMes need more active cooling than their predecessors did, or else the ASIC controller constantly makes itself thermal throttle.
You'll need hwinfo to see the temperature, as the Magician software doesn't report that sensor. It should appear in hwinfo as Drive Temperature 2.

3)850 Evo.
Perhaps it needs a driver update thru Samsung? Or maybe it's also running toasty?

4)Ram.
Gotta turn on A-XMP first.

Capt. Discombobulate may be right about the latest Nvidia driver. IDK, I'm still on 466.47.
I'm a bit of a beginner but will try my best to answer these haha.

1) I downloaded the latest AMD chipset driver after reading your post/installed it. I cloned my old SSD onto the NVME and assumed everything was good to go. Unfortunately hasn't helped the stuttering issue yet.

2) Monitored the temperature per your suggestion - it's running at a normal/average temp (55 celsius).

3) Drivers fully up to date

4) I went to the BIOS and believe I turned it on now.

5) Tried two previous driver versions after clean installs.

Not really sure what else to do, I'm going to keep tweaking in-game settings and hope I can tune it to be smooth at least. From a temp perspective everything seems to be running okay/not overheating. I ran GPU-Z but I'm not really sure what to look for. BIOS update and tweaking the voltage on the graphics card are the only things I can think of that I haven't done but as a beginner it's a bit terrifying lol.
 

Zerk2012

Titan
Ambassador
I'm a bit of a beginner but will try my best to answer these haha.

1) I downloaded the latest AMD chipset driver after reading your post/installed it. I cloned my old SSD onto the NVME and assumed everything was good to go. Unfortunately hasn't helped the stuttering issue yet.

2) Monitored the temperature per your suggestion - it's running at a normal/average temp (55 celsius).

3) Drivers fully up to date

4) I went to the BIOS and believe I turned it on now.

5) Tried two previous driver versions after clean installs.

Not really sure what else to do, I'm going to keep tweaking in-game settings and hope I can tune it to be smooth at least. From a temp perspective everything seems to be running okay/not overheating. I ran GPU-Z but I'm not really sure what to look for. BIOS update and tweaking the voltage on the graphics card are the only things I can think of that I haven't done but as a beginner it's a bit terrifying lol.
Run the benchmark again and see if it now says 3600 for the memory instead of 2133, that is probably most of your problem.
 

Mondeezy

Honorable
Jul 24, 2016
41
1
10,535
0

Mondeezy

Honorable
Jul 24, 2016
41
1
10,535
0
UBM looks a little better. Something's still off with those storage drives though.


Hold on now... I thought Memory Clock wasn't supposed to change, like EVER... can someone confirm this?

But then I forgot that me having more than one monitor plugged in causes some funky things to happen...
Yeah, that looked off to me. I took some screenshots in MSI Afterburner a few seconds after the stutters happened (attached below):

View: https://imgur.com/a/quLaCLE


Not getting enough power maybe? I'm really not sure.
 

Phaaze88

Titan
Ambassador
Not getting enough power maybe?
No, that definitely isn't it. One memory chip uses like a few watts at most, and the voltage fed to the Vram is fixed from vbios.

I don't believe the memory clock is supposed to change, but like I said, I've been on a multi monitor setup for a few years - so maybe I'm a bit out of touch.
Drivers still haven't been fixed in regards to power consumption on multi monitor. Nvidia likely gave up trying to fix that part of the drivers.
Even if I set Nvidia Control Panel power plan to Optimal, which is supposed to prioritize power saving, core clock doesn't go below 1481mhz, and memory clock doesn't go below 1377mhz.
According to Gpu-Z, board power draw is around 60w, and voltage at 0.762v.

You might have faulty Vram, IDK. We need more input from others.
 

Mondeezy

Honorable
Jul 24, 2016
41
1
10,535
0
No, that definitely isn't it. One memory chip uses like a few watts at most, and the voltage fed to the Vram is fixed from vbios.

I don't believe the memory clock is supposed to change, but like I said, I've been on a multi monitor setup for a few years - so maybe I'm a bit out of touch.
Drivers still haven't been fixed in regards to power consumption on multi monitor. Nvidia likely gave up trying to fix that part of the drivers.
Even if I set Nvidia Control Panel power plan to Optimal, which is supposed to prioritize power saving, core clock doesn't go below 1481mhz, and memory clock doesn't go below 1377mhz.
According to Gpu-Z, board power draw is around 60w, and voltage at 0.762v.

You might have faulty Vram, IDK. We need more input from others.
Yeah, fair enough. I noticed the CPU clock speeds also drop ~90% during these micro-freezes but not sure if that's the issue or a side effect.
 

Phaaze88

Titan
Ambassador
-Cpu thermal throttling.
-Cpu VRM power throttling.
-Storage thermal throttling.
-Storage has a high number of bad sectors.
-Game needs optimization.
-A driver is stuck in execution, or taking its sweet time.
That's what comes to mind, there's likely more reasons.
 
Here's the latest benchmark:

https://www.userbenchmark.com/UserRun/47224315

I also took a couple of screenshots of GPU-Z shortly after the game stuttered earlier, not sure if anything looks wrong here:

http://gpuz.techpowerup.com/21/10/20/7j9.png

Appreciate your help!
Since you just basically overclocked your RAM you really need to make sure it's stable.
Create and boot off a Memtest86 USB. Let it go through all the tests at least once (1 full pass) to confirm 99% stability.

Your scores aren't bad, really. If you want to dig further, loop unigine valley benchmark, at max settings, while HWiNFO64 (sensors only, logging on) is running in the background. Give it a few mins or until you see some bad stuttering and then quite the benchmark and go through the log file to see what's up.
 

Phaaze88

Titan
Ambassador
Cpu apparently spiked over 100C during the Cpu Test - 5800X has a max temperature average of 90C - but only like 2 applications(hwinfo & Ryzen Master) even read Ryzen cpus correctly, so that can be thrown out the window...
Gpu memory frequency actually went UP around the same time, higher than it ever did... I don't know what this means, as Gpu Boost doesn't affect memory clock. Did you overclock this?

There's that huge dip in gpu load soon after Graphics Test 1 started. Graphics test 2 had something similar at the beginning of it, but it wasn't nearly as drastic. Either way, the cpu and gpu didn't respond at all; they kept up their max boosts.
The other hard dips should be rest from fading out and transitioning to the next phase.

Do you mean VRAM clock speeds? Those change all the time.
Thanks.
It's just the Nvidia driver and my dual monitor setup then. Memory clock never goes down.
If the cpu had an iGPU, I likely could get around that.
 

Mondeezy

Honorable
Jul 24, 2016
41
1
10,535
0
OP, is anything overclocked, overvolted, or have timings adjusted? (anything being CPU, RAM, MB, or GPU)
What CPU cooler are you using?
I haven't overclocked/volted anything, I just set up the new rig a couple of days ago and have been having the micro-freeze issue whenever I try to play games (whether on low or high settings). It runs smooth the rest of the time.

Cooler is the Coolermaster Masterliquid ML360R RGB.
 
That cooler is definitely up to the task. There's something else going on here.

Did you remove the protective plastic from the bottom of the CPU block before attaching? What thermal paste did you use? Are you monitoring your Masterliquid ML360R temps?
 

ASK THE COMMUNITY