Lag spikes when opening certain apps after an incomplete update

seth · 2025-12-22 20:51:13

Dec 23 09:11:51 luthis kernel: nouveau 0000:08:00.0: NVIDIA GA102 (b72000a1)
…
Dec 23 09:12:00 luthis systemd[1]: Reached target Multi-User System.
Dec 23 09:12:00 luthis systemd[1]: Reached target Graphical Interface.
Dec 23 09:12:00 luthis systemd[1]: Startup finished in 16.570s (firmware) + 6.090s (loader) + 4.393s (kernel) + 24.700s (userspace) = 51.755s.

Time jumps backwards

Dec 23 09:11:51 luthis kernel: nouveau 0000:08:00.0: [drm] fb0: nouveaudrmfb frame buffer device
Dec 23 09:11:51 luthis systemd[1]: Finished Log2Ram.
Dec 23 09:11:51 luthis systemd[1]: Starting Journal Service...
Dec 23 09:11:51 luthis systemd-journald[742]: Collecting audit messages is disabled.
Dec 23 09:11:51 luthis systemd-journald[742]: Journal started
Dec 23 09:11:51 luthis systemd-journald[742]: Runtime Journal (/run/log/journal/54bbe2b6dca14e1899f2ef01c3c3f583) is 8M, max 3.1G, 3.1G free.
Dec 23 09:11:36 luthis systemd-modules-load[358]: Using 4 probe threads
Dec 23 09:11:36 luthis mtp-probe[388]: checking bus 5, device 7: "/sys/devices/pci0000:00/0000:00:08.1/0000:0a:00.3/usb5/5-1/5-1.4/5-1.4.2"
Dec 23 09:11:36 luthis systemd-modules-load[358]: Inserted module 'ntsync'
Dec 23 09:11:36 luthis mtp-probe[388]: bus: 5, device: 7 was not an MTP device

so I'm not sure how much we can trust those numbers

Dec 23 09:11:51 luthis dhcpcd[801]: enp4s0: waiting for carrier
Dec 23 09:11:51 luthis NetworkManager[997]: <info>  [1766434311.9671] dhcp: init: Using DHCP client 'internal'

Disable dhcpcd if you want to use NM.
It is certainly *one* stalling aspect.

Then post the journal w/ the nvidia driver.

Edit: you're also running ntpd and systemd-timesyncd, pick one.

Last edited by seth (2025-12-22 20:53:16)

luthis · 2025-12-22 21:16:43

Clearing those up. I have to do half the work over SSH, so hopefully I can still login.

Ok removed dhcpcd and disabled ntpd.

Startup finished in 16.569s (firmware) + 6.097s (loader) + 4.383s (kernel) + 21.058s (userspace) = 48.108s
graphical.target reached after 21.058s in userspace.

http://0x0.st/PzXk.txt

And now revert back to the latest nvidia (nvidia-open) drivers?

I can see there are a few other people having similar issues to me, but mine started on the 14th, a week before the latest nvidia driver change.

EDIT:

nvidia-open installed,

pacman -Q | grep nvidia
linux-firmware-nvidia 20251125-2
nvidia-open 590.48.01-3
nvidia-utils 590.48.01-1
opencl-nvidia 590.48.01-1

http://0x0.st/Pz8r.txt

Last edited by luthis (2025-12-22 21:29:13)

seth · 2025-12-22 21:34:58

And now revert back to the latest nvidia (nvidia-open) drivers?

Yes, your GPU is covered by nvidia-open. Do not try to disable GSP, though.

There's a lot of problems w/ recent nvidia drivers (580.119) flickering and general struggles to change the driver or if they have to switch to the AUR one, but if you're lucky the driver fixes it for you

If not, one thing that looks at me is log2ram - the other one is that you're using some RAID?

lsblk -f

Edit: the startup doen't look much different between nouveau and nvidia, esp. not time-wise.

Last edited by seth (2025-12-22 21:36:35)

luthis · 2025-12-22 21:42:32

I noticed the log2ram times, and cleared out some old logs that hadn't been rotated.

NAME        FSTYPE            FSVER            LABEL            UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
sda                                                                                                                 
└─sda1      linux_raid_member 1.2              luthis:RAIDArray 134bcb6f-f8dc-ceb1-3eca-b5c3f2f311bb                
  └─md127   ext4              1.0              RAID             dd7b398e-7e74-4afe-b4a2-383225ab9c51  110.2G    94% /media/RAID
sdb                                                                                                                 
├─sdb1      vfat              FAT32                             DA80-2EEB                                           
└─sdb2      ext4              1.0                               a62e9498-3da6-4070-a791-d05ac558108a                
sdc                                                                                                                 
├─sdc1      ext4              1.0                               e478a64d-ef49-42ea-8ccb-2265c3c2382e   13.2G    89% /media/ssd
└─sdc2      swap              1                                 fb727571-2925-4f93-9d25-12bde16fb852                
sdd                                                                                                                 
└─sdd1      linux_raid_member 1.2              luthis:RAIDArray 134bcb6f-f8dc-ceb1-3eca-b5c3f2f311bb                
  └─md127   ext4              1.0              RAID             dd7b398e-7e74-4afe-b4a2-383225ab9c51  110.2G    94% /media/RAID
sde         ext4              1.0              DATA             3391d243-dcc4-4ddc-98af-6deeff6a6d33                
sdf                                                                                                                 
├─sdf1                                                                                                              
└─sdf2      ext4              1.0              Series           a9a927f9-c12d-4128-acbe-cafa263957cb                
sdg         iso9660           Joliet Extension ARCH_202509      2025-09-01-16-37-02-00                              
├─sdg1      iso9660           Joliet Extension ARCH_202509      2025-09-01-16-37-02-00                              
└─sdg2      vfat              FAT32            ARCHISO_EFI      68B5-CBAE                                           
nvme0n1                                                                                                             
├─nvme0n1p1 vfat              FAT32                             4B7E-21F9                             498.8M     0% /efi
├─nvme0n1p2 swap              1                                 16d3f3f2-8a59-4dd0-b11a-1307d7a3c662                [SWAP]
└─nvme0n1p3 ext4              1.0                               a96b3354-70dd-45ed-8c6c-95171e9f1e82  445.3G    46% /var/hdd.log

I am unfortunately not able to remote desktop now for some reason, so I can't test if the issue has been resolved.. I need to drive an hour to get to my PC but I also need to get some Christmas shopping done first so... I guess I'll get on to that.

luthis · 2025-12-23 01:37:38

Unfortunately.. the issue remains.

I'm going to try dkms.

I'm running out of ideas. Issue persists in X, (although maybe lesser) so it's either hardware or a driver issue that I can't seem to resolve by downgrading.

seth · 2025-12-23 10:53:07

Errr… where's your root partition?

luthis · 2025-12-23 21:29:57

Good question!

Actually, it's here:

└─nvme0n1p3 259:3    0   929G  0 part  /var/hdd.log
                                       /

I must have missed copying the slash.

I did a bunch more testing. TL;DR, I get lag when there is load on the GPU. I turned down the graphics on Elite and was able to play it consistently, no lag (after it had loaded). If the load on the card stays small, there's no lagging out.

When I open an application, there is lag while graphics load etc, and then it smooths out when it's all loaded. I watched sensors and nvidia-smi, and .. there was nothing exciting. But, the GPU fan never rose above 50%. Not that I tested with any significant GPU usage, but it seemed to stop at 50% and never go above even briefly.

I am not sure if I could test this using a live distro, because I'd need nvidia drivers.

I'm worried that there is some part of the card that has been fried. It's a 3090.. not cheap.

luthis · 2025-12-23 23:54:58

I noticed I didn't have lib32-nvidia-utils installed, so I installed the 580xx version:

[2025-12-24T10:31:14+1300] [PACMAN] Running 'pacman -S --config /etc/pacman.conf -- multilib/lib32-nvidia-utils'
[2025-12-24T10:33:04+1300] [PACMAN] Running 'pacman -U --config /etc/pacman.conf -- /var/cache/pacman/pkg/lib32-nvidia-580xx-utils-580.119.02-1-x86_64.pkg.tar.zst'
[2025-12-24T10:34:34+1300] [ALPM] running '00-informant.hook'...
[2025-12-24T10:34:36+1300] [ALPM-SCRIPTLET] There are no unread news items
[2025-12-24T10:34:36+1300] [ALPM] transaction started
[2025-12-24T10:34:36+1300] [ALPM] installed lib32-nvidia-580xx-utils (580.119.02-1)
[2025-12-24T10:34:36+1300] [ALPM] transaction completed
[2025-12-24T10:34:36+1300] [ALPM] running '30-systemd-update.hook'...
[2025-12-24T10:34:36+1300] [PACMAN] Running 'pacman -D --asexplicit -q --config /etc/pacman.conf -- lib32-nvidia-580xx-utils'

..And then there was no lag! But also, steam wouldn't start. So, I tried another version:

[2025-12-24T12:27:27+1300] [PACMAN] Running 'pacman -R -n -s --config /etc/pacman.conf -- lib32-nvidia-580xx-utils'
[2025-12-24T12:27:28+1300] [ALPM] transaction started
[2025-12-24T12:27:29+1300] [ALPM] removed lib32-nvidia-580xx-utils (580.119.02-1)
[2025-12-24T12:27:29+1300] [ALPM] transaction completed
[2025-12-24T12:27:29+1300] [ALPM] running '30-systemd-update.hook'...
[2025-12-24T12:28:47+1300] [PACMAN] Running 'pacman -S'
[2025-12-24T12:30:22+1300] [PACMAN] Running 'pacman -R -n -s --config /etc/pacman.conf -- lib32-nvidia'
[2025-12-24T12:31:52+1300] [PACMAN] Running 'pacman -U /var/cache/pacman/pkg/lib32-nvidia-580xx-utils-580.119.02-1-x86_64.pkg.tar.zst'
[2025-12-24T12:39:02+1300] [PACMAN] Running 'pacman -U https://archive.archlinux.org/packages/l/lib32-nvidia-utils/lib32-nvidia-utils-580.95.05-1-x86_64.pkg.tar.zst'
[2025-12-24T12:39:13+1300] [ALPM] running '00-informant.hook'...
[2025-12-24T12:39:14+1300] [ALPM-SCRIPTLET] There are no unread news items
[2025-12-24T12:39:14+1300] [ALPM] transaction started
[2025-12-24T12:39:15+1300] [ALPM] installed lib32-nvidia-utils (580.95.05-1)
[2025-12-24T12:39:15+1300] [ALPM] transaction completed
[2025-12-24T12:39:15+1300] [ALPM] running '30-systemd-update.hook'...

But this lagged

So I reverted:

[2025-12-24T12:42:49+1300] [PACMAN] Running 'pacman -R -n -s --config /etc/pacman.conf -- lib32-nvidia-utils'
[2025-12-24T12:42:51+1300] [ALPM] transaction started
[2025-12-24T12:42:52+1300] [ALPM] removed lib32-nvidia-utils (580.95.05-1)
[2025-12-24T12:42:52+1300] [ALPM] transaction completed
[2025-12-24T12:42:52+1300] [ALPM] running '30-systemd-update.hook'...
[2025-12-24T12:43:33+1300] [PACMAN] Running 'pacman -U --config /etc/pacman.conf -- /var/cache/pacman/pkg/lib32-nvidia-580xx-utils-580.119.02-1-x86_64.pkg.tar.zst'
[2025-12-24T12:43:34+1300] [ALPM] running '00-informant.hook'...
[2025-12-24T12:43:34+1300] [ALPM-SCRIPTLET] There are no unread news items
[2025-12-24T12:43:34+1300] [ALPM] transaction started
[2025-12-24T12:43:34+1300] [ALPM] installed lib32-nvidia-580xx-utils (580.119.02-1)
[2025-12-24T12:43:34+1300] [ALPM] transaction completed
[2025-12-24T12:43:34+1300] [ALPM] running '30-systemd-update.hook'...
[2025-12-24T12:43:34+1300] [PACMAN] Running 'pacman -D --asexplicit -q --config /etc/pacman.conf -- lib32-nvidia-580xx-utils'

But.. we are still stuck at the lagging. Somehow.

luthis · 2025-12-24 00:57:18

I have managed to fix it .. again. Hopefully permanently.

I used the previous package in my local cache for nvidia-utils from october that matches the version of the other packages:

lib32-nvidia-utils 580.95.05-1

and for some reason, this works!

Here's the other installed packages from the repo:

linux-firmware-nvidia 20251021-1
nvidia-dkms 580.95.05-1
nvidia-utils 580.95.05-1
opencl-nvidia 580.95.05-1

All downgraded, the latest packages don't seem to work at all for me.

I have no idea why.

Last edited by luthis (2025-12-24 00:57:37)

seth · 2025-12-24 16:43:36

W/ the flickering of the 585.119 drivers the lags might have the cause rooted in handling of the vblank signal.
The nvidia-open driver (with properly updated lib32-nvidia-utils) did also not work?

luthis · 2026-01-04 09:40:24

Just tested.. I updated everything and had severe issues. Couldn't start niri or awesome.

Rolled back to the drivers listed earlier, but this now gives me the lag, although not as bad perhaps?

seth · 2026-01-04 12:36:35

Couldn't start niri or awesome.

Please don't paraphrase, https://bbs.archlinux.org/viewtopic.php?id=57855
Do you have error messages/backtraces etc?

luthis · Yesterday 08:37:06

Hi,

Sorry I've been busy and hadn't had a chance to troubleshoot more; After the second reboot after downgrading, it was all working again. I was able to play a bunch of VR and used sunshine to play remotely all week. Zero issues.

I had to update today, and the issues returned. I rolled back again, didn't resolve. Since the issue happens with Mint as well and the only common factor was Grub, I moved to systemd-boot instead. Same issues.

I booted to a live EndeavourOS that comes with the latest nvidia drivers so I can eliminate my entire OS from the problem. *same issue.*

So, I can only conclude that there is some hardware fault or BIOS issue with my PC... except I was gaming all week and today (in VR + flat) with *ZERO ISSUES* until I updated.

I have no idea what to do here. I can reset my BIOS?

---

Ok I updated the BIOS firmware... no luck.

Here's the xorg logs:

http://0x0.st/PXtY.txt

Journal logs:

http://0x0.st/PXv9.txt

I also unplugged everything unnecessary, switched to HDMI from DP..

I've eliminated all the software, so I must assume it's hardware. But .. it was working until I updated...

Last edited by luthis (Yesterday 09:53:28)

seth · Yesterday 16:13:05

But .. it was working until I updated...

… and rebooted?

xrandr --verbose

resp.

for OUT in /sys/class/drm/card*; do echo $OUT; edid-decode $OUT/edid; echo "================="; done

You'll need https://archlinux.org/packages/extra/x86_64/v4l-utils/

luthis · Yesterday 21:47:11

Perhaps yes, I don't recall if I rebooted after I got it working last week.

I also tested an EndeavourOS from 2024, same issue.

xrandr:

http://0x0.st/PXlt.txt

second one:

http://0x0.st/PXlv.txt

On a positive note, I'm back to consistent sub-30-second boot times.

When I checked the bootplot svg, it loaded in firefox, and I got the lag for a few seconds. Then also when I closed firefox, it lags again, until stabilising. So it's both during loading / processing / and unloading.

That leads me to suspect a ram issue. I might try removing one stick and swapping them round. But surely if it was hardware like that, it would be having issues all the time, not magically work fine for a week.

---

running stress-ng, it's using 100% of one core, 50% of ram.. and no lag spikes.

stress-ng --vm 1 --vm-bytes 50% --vm-method all --verify -t 10m
stress-ng: info:  [11774] setting to a 10 mins run per stressor
stress-ng: info:  [11774] dispatching hogs: 1 vm
stress-ng: info:  [11775] vm: using 26.86G per stressor instance (total 26.86G of 53.73G available memory)
stress-ng: info:  [11774] skipped: 0
stress-ng: info:  [11774] passed: 1: vm (1)
stress-ng: info:  [11774] failed: 0
stress-ng: info:  [11774] metrics untrustworthy: 0
stress-ng: info:  [11774] successful run completed in 10 mins, 5.69 secs


stress-ng --vm 1 --vm-bytes 90% --vm-method all --verify -t 10m
stress-ng: info:  [18144] setting to a 10 mins run per stressor
stress-ng: info:  [18144] dispatching hogs: 1 vm
stress-ng: info:  [18145] vm: using 37.35G per stressor instance (total 37.35G of 41.50G available memory)
stress-ng: info:  [18144] skipped: 0
stress-ng: info:  [18144] passed: 1: vm (1)
stress-ng: info:  [18144] failed: 0
stress-ng: info:  [18144] metrics untrustworthy: 0
stress-ng: info:  [18144] successful run completed in 10 mins, 5.56 secs

Last edited by luthis (Today 00:00:56)

luthis · Today 05:24:46

I've been doing a ton of testing. Nothing has helped yet, except:

sudo nvidia-smi -pm 1 
sudo nvidia-smi -lgc 210,210

This makes the lag when opening firefox or steam slightly better, and reduces the lag spikes in game. Has not solved it at all but definitely there is a difference

Arch Linux

#26 2025-12-22 20:51:13

Re: Lag spikes when opening certain apps after an incomplete update

#27 2025-12-22 21:16:43

Re: Lag spikes when opening certain apps after an incomplete update

#28 2025-12-22 21:34:58

Re: Lag spikes when opening certain apps after an incomplete update

#29 2025-12-22 21:42:32

Re: Lag spikes when opening certain apps after an incomplete update

#30 2025-12-23 01:37:38

Re: Lag spikes when opening certain apps after an incomplete update

#31 2025-12-23 10:53:07

Re: Lag spikes when opening certain apps after an incomplete update

#32 2025-12-23 21:29:57

Re: Lag spikes when opening certain apps after an incomplete update

#33 2025-12-23 23:54:58

Re: Lag spikes when opening certain apps after an incomplete update

#34 2025-12-24 00:57:18

Re: Lag spikes when opening certain apps after an incomplete update

#35 2025-12-24 16:43:36

Re: Lag spikes when opening certain apps after an incomplete update

#36 2026-01-04 09:40:24

Re: Lag spikes when opening certain apps after an incomplete update

#37 2026-01-04 12:36:35

Re: Lag spikes when opening certain apps after an incomplete update

#38 Yesterday 08:37:06

Re: Lag spikes when opening certain apps after an incomplete update

#39 Yesterday 16:13:05

Re: Lag spikes when opening certain apps after an incomplete update

#40 Yesterday 21:47:11

Re: Lag spikes when opening certain apps after an incomplete update

#41 Today 05:24:46

Re: Lag spikes when opening certain apps after an incomplete update

Board footer