[SOLVED] High latency issues on new computer

Nov 1, 2019
11
1
15
Hi all,

I finished building my new gaming computer recently and facing serious audio (popping, crackling, stuttering) issues as soon I start a game. These issues are appearing in both, the onboard sound device and the dedicated gameDAC from Steelseries. At least the gameDAC and its headset are fine when used on another computer.

These issues are most present when a certain load is on the computer like gaming. Watching youtube videos does not show issues.

Last week I tried many (unsuccessful) fixing attemps, thats why I'm reaching out for help.

A short list of the important things done so far:

- uninstall devices, update drivers, bios, chipset, etc.

- power plan set to high performance mode

- All connections/wires checked and hardware repositioned

- uninstalled all software related to audio, just re-installed Steelseries Engine for my DAC and headset

- deactivated wifi and onboard sound in bios

- Windows, drivers, nvidia software all on their latest version

I was running latencymon while gaming, as these issues are very present then. The results can be found via this link:

Log File:
https://drive.google.com/open?id=1lgKKXx3VGcO7J63H3-9_anVEmmjJ4IcS

CPU:
https://drive.google.com/open?id=1tWruWnyJTP5x7D85CgRqRl2YKgpr0IEg

Driver:
https://drive.google.com/open?id=1xOUEdbo7XdBALSVgsb6ahEhQHFbKtjmp

Process:
https://drive.google.com/open?id=1rOaPStL88CkgMgpUek-5S9x0-KWBuoES

My computer:

  • Windows 10 Pro 64bit 1903 (18362.449)
  • AMD Ryzen 3900x
  • ASUS ROG 2080 ti
  • ASUS ROG Crosshair VIII Hero (wifi)
  • Seasonic Focus Plus Platinum (750W)
  • 2x Samsung Evo Plus M.2
  • 2x 16gb G.skill Tridenz Z Neo (@3600mhz, 16-16-16-36)

Thank you in advance for any provided help.
 
Solution
Yes, 1.36v is common on some boards. It's not a problem and WELL within the tolerance for voltage on DDR3.

Try this. First, make sure they are in the correct slots, A2 and B2, which are the second and fourth slots to the right of the CPU socket. Then, do a hard reset of the BIOS as follows:


BIOS Hard Reset procedure

Power off the unit, switch the PSU off and unplug the PSU cord from either the wall or the power supply.

Remove the motherboard CMOS battery for five minutes. In some cases it may be necessary to remove the graphics card to access the CMOS battery.

During that five minutes, press the power button on the case for 30 seconds. After the five minutes is up, reinstall the CMOS battery making sure to...
Do you have the latest motherboard BIOS version installed?

Are your memory modules in the A2 and B2 slots?

Have you pulled the CPU to check for a bent pin or two?

Pulled the motherboard to verify that there are no standoffs where they don't belong?

Also, you power plan. Make SURE you have the latest chipset version for your motherboard, FROM the AMD website, not the board manufacturer. And that you have it set to balanced, not performance. I've seen this go both ways, with some users saying they issues with one and not the other so trial and error might rule here.

Also, just being on the "latest version" almost means nothing these days. Doing a clean install of the Nvidia drivers is imperative AND you might want to download the DDU and the latest Nvidia driver package, then boot to safe mode, unplug your internet connection, use the DDU to remove the current drivers, restart, install the latest Nvidia drivers (Standard) and THEN reconnect your internet connection so you don't run into any DCH driver version issues.
 
  • Like
Reactions: Reaven-X
Thank you very much. Yes i have installed the latest bios for my motherboard before Win 10 was installed. Chipset was updated from AMD Website, but version didn't change, so it seems Asus had already the latest one distributed via armory crate software.

I uninstalled nvidia driver in "apps", did a restart and a fresh install via nvidia experience software. Gonna try DDU and w/o ethernet.

Stand-offs were fine, checked them carefully before I installed the motherboard, the same counts for the CPU, had checked the Pins before I carefully placed it into the socket.

Will give an update.
 
Looks to me like it's Nvidia driver related. Reinstall the driver using the Custom > Clean Install option. Don't let it install any of the Nvidia audio crap.

Actually you're getting hard page faults....that points to bad RAM. You try memtest? Any system crashes?
 
Thank you, I downloaded DDU and the latest standard (not DCH) driver for my 2080 ti.

I booted windows in safe mode, unplugged ethernet and deleted screen and gpu driver with DDU. Geforce experience has been uninstalled too.

I restarted win 10 in normal mode (still without ethernet) and installed the newest driver manually w/ clean install option. Only thing installed next to the driver was NVIDIA PhysX software. But no USB C or HD Audio drivers, don't need them.

Unfortunately I still get these audio issues, specially present when a higher load is applied on the computer, like in cut scenes (CoD:MW). Seen as increase in CPU temp.
CPU temp while gaming stays between 45-55°C, GPU 60-65°C and VRM 30-35°C. I think thats pretty ok and should not trottle anything.

I checked the stand-off's again (just to be sure) and there is one for each screw hole in the motherboard, no cables nor anything else is touching the back of the board. It looks perfect if I may say so. :sweatsmile:

I downloaded memtest86 and let it ran with 13 different test algorithms for 2 full cycles (3 hours). 0 errors were found, so I assume my RAM runs fine @3600mhz. DIMM's are installed in A2 & B2.

You can find a picture of the results -> here


I ran latencymon again while CoD:MW was running.

Latencymon

Latencymon detailed

Drivers

Processes

CPU


I hope these information are usefull, otherwise please let me know what else I can provide.
Many thanks for helping me. :)


EDIT: Just as additional information, it is not CoD:MW related, as I get these issues also in another game.
Onboard sound and USB DAC showing the same issue, first one was then disabled in Bios as I don't need it, the same counts for wifi and bluetooth.
 
Last edited:
You using a wired ethernet connection or wireless? This can be related to network drivers. Worth trying to find new or different drivers. You can also try going into the Device Manager and uninstalling the devices and then telling it to scan for new hardware to reinstall them.
 
  • Like
Reactions: Reaven-X
Ethernet, I uninstalled my intel lan adapter and updated its driver to the newest one from intel website. Did not fix the issue. Then I disabled this adapter in bios and enabled the second one on my motherboard (realtek 2.5gb), but this also didn't resolve my issue. :/

Is there a way to reinstall DX12? I understand that dxgkrnl.sys (showing highest latency) might be the cause of my nightmares.
 
I have to agree that at this point, a clean install might be the best option as suggested by Anort3. At least you can eliminate that as being a potential cause of the issue, if it hasn't already been done sometime recently. Keep in mind, refresh, reset, upgrade installations, spring and fall updates, these are not clean installs.
 
  • Like
Reactions: Reaven-X
I installed Win 10 on this new computer 2 weeks ago using a USB stick. There was never an older version installed.

Do I loose files when I do a clean reinstall? Because I just googled how to do a clean reinstall and found there is an option to keep files.

Further, I have 2 M.2 SSD's installed. On (C: ) is the OS, game launchers and files, the other one (D: ) is the storage for my games. Do I loose D: when reinstalling Windows 10?

Sorry, never did that with this OS.
 
Just tried one last thing before wiping my computer. I disabled DOCP (XMP) and ran my Ram at ddr4 stock 2133mhz.

And guess what happened....No issues!

Strange that memtest showed 0 errors.

I kept docp off and added manually the settings according to XMP profile of my RAM (what I paid for) but audio popps / stutters was immediately back.
 
I thought those hard page fault errors pointed to the RAM. I'm not sure what to tell you as far as getting the RAM speed you paid for with no audio errors though. I'd try memtest again this time for 10-12 hours and see if it shows any errors.
 
  • Like
Reactions: Reaven-X
If this was an Intel platform, I'd say that lowering the VCCIO and VCCSA voltages might do the trick. Auto voltage on many Intel platforms can sometimes cause audio problems when XMP profiles above 2666mhz are used. I've seen it a few times. I still think those are board issues though, even in those cases.

TBH, I don't know what Ryzen settings are equivalent to VCCIO and VCCSA that you could potentially look at lowering the voltage on, since we know that practically all auto voltages for all platforms, are raised excessively to err on the side of stability.

Example: https://forums.tomshardware.com/threads/ram-xmp-profile-audio-issues.2804618/

https://forums.anandtech.com/thread...o-onboard-audio-and-usb.2534412/post-39424427


Having the latest motherboard BIOS should be a no brainer here as well.
 
  • Like
Reactions: Reaven-X
Many thanks to both, it ran stable for about 2h of gaming @3200mhz CL14 1.35v, which is very similar to 3600 CL16 1.35v from XMP profile as far I understand.

I'm not very familar with OC and especially AMD, as it is my first CPU from this company. I'll work trough it the next days and might do some additional changes in setting to find the sweatspot for my system.

Really happy atm, first time since 3 weeks that I actually have enjoyed my new rig.
 
If you have changed the memory timings, you need to do some stability testing to ensure that the system is FULLY stable. You can find the test procedures here:

 
  • Like
Reactions: Reaven-X
Thanks!

I retested it with Memtest86 last night (4 cycles, each with all Tests, ~6h total) and got 5 errors.
So I assume my RAM has indeed a problem?

I got in touch with someone who has the identical set up of Motherboard, CPU and RAM. His system runs without issues with XMP enabled.

Btw, one more question. Is it normal that Bios and monitoring software showing my RAM at 1.36v, even though XMP or I (manually) set Voltage to 1.35?
 
Yes, 1.36v is common on some boards. It's not a problem and WELL within the tolerance for voltage on DDR3.

Try this. First, make sure they are in the correct slots, A2 and B2, which are the second and fourth slots to the right of the CPU socket. Then, do a hard reset of the BIOS as follows:


BIOS Hard Reset procedure

Power off the unit, switch the PSU off and unplug the PSU cord from either the wall or the power supply.

Remove the motherboard CMOS battery for five minutes. In some cases it may be necessary to remove the graphics card to access the CMOS battery.

During that five minutes, press the power button on the case for 30 seconds. After the five minutes is up, reinstall the CMOS battery making sure to insert it with the correct side up just as it came out.

If you had to remove the graphics card you can now reinstall it, but remember to reconnect your power cables if there were any attached to it as well as your display cable.

Now, plug the power supply cable back in, switch the PSU back on and power up the system. It should display the POST screen and the options to enter CMOS/BIOS setup. Enter the bios setup program and reconfigure the boot settings for either the Windows boot manager or for legacy systems, the drive your OS is installed on if necessary.

Save settings and exit. If the system will POST and boot then you can move forward from there including going back into the bios and configuring any other custom settings you may need to configure such as custom fan profile settings or other specific settings you may have previously had configured that were wiped out by resetting the CMOS, but do NOT set the memory back to the XMP profile. Leave the memory at whatever it defaults to.

Save settings, exit the BIOS and make sure it will POST at the default settings.

Run Memtest86 again with the memory at the non-XMP default settings and see if you get the same errors again. If you do, RMA the memory. If you do not, set the memory to the XMP configuration again BUT THIS TIME go into the memory settings and check to see whether the command rate is set to 1T or 2T. If it is set to 1T, change it to 2T and test again. If it is already set to 2T then don't change it. We will go from there.
 
  • Like
Reactions: Reaven-X
Solution
Thank you again. Much appreciated!
Sure, I can run another test with 2T the next days.

But for my understanding, I just read following article which included a paragraph about 1T and 2T:

Best Memory Settings for AMD's 3000 CPUs

I found out my RAM (F4-3600C16D-32GTZN) 2x 16GB is dual-ranked. One DIMM is in A2 the other one in B2.
Based on the mentioned article, they say 2T is important when two “double-sided” or four “single-sided” DIMMs are used. As mine is dual-ranked, shouldn't 2T be set in Bios by its XMP profile? I checked in CPU-Z and in tab "Memory" the command rate is 1T. In tab "SPD" are only the first 5 timings, but nothing mentioned about command rate.

I always thought XMP contains all needed information, but regarding this article, it would need 2T instead of 1T?

I hope I understood it the correct way, please let me know if I'm wrong.
 
Last edited:
XMP contains MANY more timings than just the main timings. XMP usually contains primary, secondary AND tertiary timings, although some are usually left up to the discretion of the motherboard to automatically configure. Look in the advanced memory settings for your motherboard in the BIOS. If it is a board that allows changes to memory timings then it will be there, probably in the secondary settings.

Gear down could be a factor as well. There is a lot of good information on your board in this thread.

https://rog.asus.com/forum/showthread.php?101617-Crosshair-VII-Hero-Essential-Info-Thread/page5
 
  • Like
Reactions: Reaven-X
Update:

I realized that SOC voltage on "auto" takes 1.1v, which should be normal for 3rd gen ryzen, but CLDO VDDG took also 1.1v on "auto", but should be -0.1v lower than SOC voltage. After reducing it manually to 1.0v, everything seems to run just fine at 3600mhz CL16. Will run memtest and prime again, as soon I find some time, but a 3h session of gaming passed w/o issues.

Looks like a bug of the "auto" setting in Bios.