Massive stuttering, audio popping/cracking in games

[Gmr_Chick]

I wasn't quite sure where to post this. Basically, I've been encountering massive stuttering in games recently. The audio crackles and pops, my USB devices act up (wireless headphones disconnect, among other things) and I've had several BSODs already recently in the past couple weeks or so. It's doing it right now as I type this, in fact. It's very aggravating. I've updated my chipset drivers, audio drivers, LAN drivers, BIOS, GPU, but still the problem continues. Earlier tonight, I managed to fire up LatencyMon just as the issue started and got this:

And, as I've still got the program running, the readings look like this now:

I literally don't know what more to do. I've thought maybe the ethernet controller is bad on my board, as its the notorious Intel i225V, but I'm not sure buying a whole new board would help fix the problem. Any help on this issue would be greatly appreciated!

 

[eidairaman1]

Get blue screen viewer and post up your bsods and also event viewer for diagnostics, also while you are at it post up hwinfo64 screenshots with sensors only-post everything from it here

 

[Gmr_Chick]

And screencaps of my HWINFO:

 

[tabascosauz]

@Gmr_Chick I was gonna say, that all still sounds like uncore stability like on AM4.

Problem is, things have changed a bit since AM4. Before, you would have just bumped up VSOC a bit and thrown in the towel and bought a different board if that didn't help. VSOC has a bit of a complicated relationship now and Fabric doesn't like it not only when VSOC is insufficient, but also when it's too high. 1.25V is functionally about as high as you ever want to go now (yes, I know the hard limit is 1.3V).

Unfortunately I haven't ever encountered USB issues yet on AM5 so I can't advise from experience.

When was the last time you clean installed windows?

And all the usual recommendations, make sure CPU is fully stable (even stock) and thoroughly test RAM/Fabric/UCLK (as opposed to AM4 where UCLK really wasn't a standalone thing).

I have had some disconnection issues from i225 (I have Killer E3100, same thing), but installing the actual latest driver from Intel website seems to have stopped all that, as opposed to the default one that Win Update fetches. I don't think that would cause this behaviour you're seeing.

 

[Gmr_Chick]

@Gmr_Chick I was gonna say, that all still sounds like uncore stability like on AM4.

Problem is, things have changed a bit since AM4. Before, you would have just bumped up VSOC a bit and thrown in the towel and bought a different board if that didn't help. VSOC has a bit of a complicated relationship now and Fabric doesn't like it not only when VSOC is insufficient, but also when it's too high. 1.25V is functionally about as high as you ever want to go now (yes, I know the hard limit is 1.3V).

Unfortunately I haven't ever encountered USB issues yet on AM5 so I can't advise from experience.

When was the last time you clean installed windows?

And all the usual recommendations, make sure CPU is fully stable (even stock) and thoroughly test RAM/Fabric/UCLK (as opposed to AM4 where UCLK really wasn't a standalone thing).

I have had some disconnection issues from i225 (I have Killer E3100, same thing), but installing the actual latest driver from Intel website seems to have stopped all that, as opposed to the default one that Win Update fetches. I don't think that would cause this behaviour you're seeing.

Is there a particular setting I should stick to for UCLK? VSOC?

Admittedly, it's been a while since I did a clean install of Windows.

 

[tabascosauz]

Is there a particular setting I should stick to for UCLK? VSOC?

Admittedly, it's been a while since I did a clean install of Windows.

No, UCLK is memory controller freq. It and FCLK are fine as they are, I just thought I'd remind you to do some memtesting while you're at it as UCLK is part of the stability equation now. VSOC is whatever lowest value allows you to consistently past memtesting without any other adverse effects.

I think popping is more likely to be a hardware issue, but the rest of the symptoms are not impossible for Windows to cause. Even on a perfectly stable system I wouldn't trust normal channel Windows 10 and 11 to run itself properly past 2 years, it occasionally just grenades things naturally through bad updates.

 

[DaemonForce]

My first thought was check your vSOC. You want something marginally better than this:

Even on a perfectly stable system I wouldn't trust normal channel Windows 10 and 11 to run itself properly past 2 years, it occasionally just grenades things naturally through bad updates.

Lately it's gotten down to as bad as every 2 months. Stay off the regular Win10 channels unless you're okay with obliterating it and starting over every 2-3 months. The lack of store apps that will actually behave is completely insufferable. Can't even run a calculator.

 

[Gmr_Chick]

No, UCLK is memory controller freq. It and FCLK are fine as they are, I just thought I'd remind you to do some memtesting while you're at it as UCLK is part of the stability equation now. VSOC is whatever lowest value allows you to consistently past memtesting without any other adverse effects.

I think popping is more likely to be a hardware issue, but the rest of the symptoms are not impossible for Windows to cause. Even on a perfectly stable system I wouldn't trust normal channel Windows 10 and 11 to run itself properly past 2 years, it occasionally just grenades things naturally through bad updates.

I was afraid it might be a hardware issue Would it matter if I bought my motherboard as a refurb (via an ebay trusted seller) and GPU used?

 

[tabascosauz]

I was afraid it might be a hardware issue Would it matter if I bought my motherboard as a refurb (via an ebay trusted seller) and GPU used?

Maybe. I meant more reliant on just the motherboard model in general, but yeah that doesn't inspire confidence.

If you have any ol' spare SATA drive lying around, you could put a stripped down install of Windows on there, disconnect your usual OS drive and test for a bit. Symptoms showing up there should be a clear indication it's not software-related.

 

[ShrimpBrime]

You have 24gb of available memory, but using 34% of available virtual memory. Seems a little bit odd. Maybe something to investigate.

 

[OkieDan]

I have a similar problem and ran across this which is about a different motherboard than I have but I think is same as yours. After removing all but one nvme, things got better but still not perfect. You might try doing the same and trying different slots for the one nvme you keep installed while troubleshooting.

 

[Gmr_Chick]

So I ran OCCT all night last night and no errors were detected during the overnight run when testing CPU+RAM. Few hours later, I ran the extreme variant for an hour. Again, no errors. Then I went into BIOS and returned everything to stock, with only the EXPO 1 setting active for my RAM kit, botted into Windows, ran OCCT again for 2 hours, no errors.

Unfortunately, I don't have a spare AM5 board or another kit of DDR5 RAM lying around, so I can't say for certain it's a problem with my current board + RAM combo. Not quite sure what else to do outside of doing a clean install of Windows.

 

[tabascosauz]

So I ran OCCT all night last night and no errors were detected during the overnight run when testing CPU+RAM. Few hours later, I ran the extreme variant for an hour. Again, no errors. Then I went into BIOS and returned everything to stock, with only the EXPO 1 setting active for my RAM kit, botted into Windows, ran OCCT again for 2 hours, no errors.

Unfortunately, I don't have a spare AM5 board or another kit of DDR5 RAM lying around, so I can't say for certain it's a problem with my current board + RAM combo. Not quite sure what else to do outside of doing a clean install of Windows.

Does OCCT even test RAM well? I only used it to stress cores. Run some TM5 for memory.

The audio issues really shouldn't have anything to do with your choice of RAM though.

A clean install really does help though:

 

[Gmr_Chick]

I probably will do a clean install once I find my thumbdrive that has Windows installed on it. At the moment though, I'm keeping my eyes on the weather and preparing. We're supposed to be getting some pretty significant rainfall this week from the remnants of the hurricane/tropical storm that just went through Florida.

Once it passes through though, into the garage I go, hunting for that thumbdrive!

 

[RJARRRPCGP]

Were you gaming when you got the "KMODE_EXCEPTION_NOT_HANDLED" in wof.sys crash?

I got that stop code before (even when it was with Windows 2000) when my AMD CPU (that happened to be on socket 462, so long ago, back in 2007) was too hot.

 

[Gmr_Chick]

Were you gaming when you got the "KMODE_EXCEPTION_NOT_HANDLED" in wof.sys crash?

I got that stop code before (even when it was with Windows 2000) when my AMD CPU (that happened to be on socket 462, so long ago, back in 2007) was too hot.

I was gaming yes. Thus far, the crashing and other issues have mainly occurred while playing Star Wars Battlefront II (2017). But it's also occurred while playing LEGO Star Wars: The Skywalker Saga. I haven't tested with The Sims 4 yet.

 

[Bawdyknocker]

I jus happened to come across this thread, and helped my GF with a similar issues with the audio crackling a couple weeks ago. now, admittedly, she is on AM4, but the issue seems related?

i read somewhere where the crackling is related to the buffer underruns in the RAM bus, and the crackling comes from some kinda crosstalk... have you tried reducing the XMP freq? it will take time to step through them and test for stability, rebooting, etc, but i'm sure you're used to that by now ^_^**

I was trying to systematically check the RAM frequencies, the sticks themselves, and the mobo in order to isolate which - if any - of those could be having the issue. and yet, the fix seemed to be to reduce the frequency setting by just one step on her XMP profile... her DDR4 is spec'd for 3200MHz, but reducing the XMP from 3200 down to 3166MHz seemed to completely eliminate the issue.

I cant imagine *that* could be the only thing affecting system wide stability, but seems plausible? I sure hopw a simple thing like this works for you, OP ^,..,^ best of luck!

 

[Gmr_Chick]

Ok, so last night I was playing Battlefront II when the issues I've discussed started happening midway through the match -- I busted my ass to record it, and now I have tangible proof. Maybe the video can help to explain what exactly is happening.


Battlefront II Issues

Had to post it on my Drive because the file is very large. And I apologize for the audible mouse clicking and crappy gameplay. I couldn't really play the way I usually do with the constant stuttering.

 

[OkieDan]

Just out of curiosity, is your system really sluggish when in BIOS setup?

 

[pavle]

Ok, so last night I was playing Battlefront II when the issues I've discussed started happening midway through the match -- I busted my ass to record it, and now I have tangible proof. Maybe the video can help to explain what exactly is happening.


Battlefront II Issues

Had to post it on my Drive because the file is very large. And I apologize for the audible mouse clicking and crappy gameplay. I couldn't really play the way I usually do with the constant stuttering.

Please try lowering your RAM speed to 5200 and test with such to see if it helps.

 

[RJARRRPCGP]

I was gaming yes. Thus far, the crashing and other issues have mainly occurred while playing Star Wars Battlefront II (2017). But it's also occurred while playing LEGO Star Wars: The Skywalker Saga. I haven't tested with The Sims 4 yet.

Darn! Looks like unstable CPU. Same stop code Windows 2000 gave me on my Barton XP 3000+ AQZFA 03xx (2003 date code) build back in 2007, after the air conditioner duct came off, because of the duct tape. I was OC'ing my Barton, and I wasn't even running anything that was taxing on an Athlon XP! Heck, it would be playable on a Pentium II, very likely!

Unless it was the RAM getting too hot, but in my case, the same cooling was for the CPU.

Lately it's gotten down to as bad as every 2 months. Stay off the regular Win10 channels unless you're okay with obliterating it and starting over every 2-3 months. The lack of store apps that will actually behave is completely insufferable. Can't even run a calculator.

Yikes! Even with my bad luck of finding buggy things in Windows 11, it was never that bad, IMX.

 

[Gmr_Chick]

Just out of curiosity, is your system really sluggish when in BIOS setup?

A bit yes. It's usually just faster to navigate through them with my keyboard than with my mouse. I just assumed the slowness was from Gigabyte's BIOS though, because I never had slowness in navigating ASUS BIOS when I briefly had an X670E Hero board.

 

[RJARRRPCGP]

A bit yes. It's usually just faster to navigate through them with my keyboard than with my mouse. I just assumed the slowness was from Gigabyte's BIOS though, because I never had slowness in navigating ASUS BIOS when I briefly had an X670E Hero board.

IIRC, I had the issue where it was slow, with the mouse pointer in the UEFI-BIOS setup on a Gigabyte motherboard before, but, also saw at least similar with others.

On legacy-BIOS, running a new video card, especially later than a Radeon R9 series, and GeForce GTX 980 series, increases the chance of that issue, along with the CMOS settings only being saved to RAM, so if you unplug the PSU, the CMOS is reset! It happens with socket 1366 motherboards. (But, not all of them, apparently. Could some 1366 motherboards be UEFI?!)

That's what happens with my XFX RX 580 on my Asus P6T6 WS Revolution.

 

[Ruru]

Just out of curiosity, is your system really sluggish when in BIOS setup?

Now this sounds familiar, I had this exact issue with Crosshair VII Hero. It went slower and slower until a single keypress took like ~5min to register. Also the clock didn't work realtime.

So I just had to rush with setting things up etc.

IIRC, I had the issue where it was slow, with the mouse pointer in the UEFI-BIOS setup on a Gigabyte motherboard before, but, also saw at least similar with others.

On legacy-BIOS, running a new video card, especially later than a Radeon R9 series, and GeForce GTX 980 series, increases the chance of that issue, along with the CMOS settings only being saved to RAM, so if you unplug the PSU, the CMOS is reset! It happens with socket 1366 motherboards. (But, not all of them, apparently. Could some 1366 motherboards be UEFI?!)

That's what happens with my XFX RX 580 on my Asus P6T6 WS Revolution.

I'm not 100% sure but I think that some of the later ones are.

 

[RJARRRPCGP]

Forgot to add that with the Athlon XP build, the RAM was cooled with the same intake as the CPU.

RIP 2004 Asus A7N8X-X (KZG caps suddenly were bulging and leaking in the mid-2010s!) (I got it in late-2005, IIRC)

Now this sounds familiar, I had this exact issue with Crosshair VII Hero. It went slower and slower until a single keypress took like ~5min to register. Also the clock didn't work realtime.

What did you think happened? That's nuts, even with heavy CPU thermal throttling, due to a bad heatsink mount!