Steam games not launching; PC performance tanks when trying

Alrighty, the sync said I was up to date and there was “nothing to do” for the second part of the first command.

The second command gave me this output:

> Removing video-hybrid-amd-nvidia-prime...
Sourcing /etc/mhwd-x86_64.conf
Has lib32 support: true
Sourcing /var/lib/mhwd/local/pci/video-hybrid-amd-nvidia-prime/MHWDCONFIG
Processing classid: 0300
Sourcing /var/lib/mhwd/scripts/include/0300
checking dependencies...
error: failed to prepare transaction (could not satisfy dependencies)
:: removing nvidia-utils breaks dependency 'nvidia-utils' required by nvidia-settings
:: removing libxnvctrl breaks dependency 'libxnvctrl' required by nvidia-settings
Error: pacman failed!
Error: script failed!

Which I assume is not supposed to happen…

As for the monitors; yes I use multiple monitors. I have a 4k monitor as my primary display, and 2 1080p monitors. All are plugged into the dGPU. I’m not sure why that would cause a problem, but if you know of some reason that it could be the case, that’s a lead of some sort.

I know that my combo of AMD CPU with integrated graphics + Nvidia GPU is not the most common, as AMD only made the integrated graphics standard as of the 7000 series chips, and Nvidia doesn’t have a great reputation with the Linux community at-large as much as I’ve ever heard. I assume that isn’t the problem though, since there is explicitly a package for amd-nvidia pairing.

I could certainly try following the steps there to make sure that my installation is working properly, but I will avoid getting too deep into that until resolving the other ideas that have been put forward; so as to not get things mixed up.

It looks like the maintainers made a mistake.

They added nvidia-settings to all nvidia profiles (including hybrid-amd-nvidia-470xx-prime), except video-hybrid-amd-nvidia-prime.

(10 months ago :no_mouth:)

For now I suggest removing nvidia-settings first (using dd to skip dependency checks);

sudo pacman -Rdd nvidia-settings

Then do the mhwd thing again

sudo mhwd -f -i pci video-hybrid-amd-nvidia-prime

At which point you can reinstall nvidia-settings;

sudo pacman -Syu nvidia-settings

@Yochanan being pinged for this package issue.

Fixed with mhwd-db 0.6.5-29 coming along shortly.

Oh, nicely spotted. The instructions you gave did let me update the driver package. That being said, the issue I was having with games not launching is still present. The driver is up-to-date now, so unless I misunderstood and there is still a driver issue; there is some other issue at hand.

I saw in another forum thread that someone recommended trying:

steam 2>&1  | tee /tmp/dumps

…to generate a log, and I made a new one of that, with the log of trying to launch a game through Steam and it not working. Is that something I should post here, or is that not a useful piece of information? I have tried looking at it before, but couldn’t really figure out what the actual problem was, and the red warning/error text didn’t help me when searched.

That being said, I appreciate your help thus far :slight_smile:

Hi,

fresh stable up date :

[nls@lap ~]$ date -u
sam. 27 janv. 2024 11:21:40 UTC
[nls@lap ~]$ sudo pacman -Syu
[sudo] password for nls: 
:: Synchronising package databases...
 core                                                                          145,4 KiB  2,54 MiB/s 00:00 [###############################################################] 100%
 extra                                                                           8,6 MiB  31,1 MiB/s 00:00 [###############################################################] 100%
 community                                                                      29,0   B  4,72 KiB/s 00:00 [###############################################################] 100%
 multilib                                                                      143,6 KiB  20,0 MiB/s 00:00 [###############################################################] 100%
:: Starting full system upgrade...
resolving dependencies...
looking for conflicting packages...
warning: dependency cycle detected:
warning: mhwd-db will be installed before its mhwd dependency
warning: dependency cycle detected:
warning: mhwd-db will be installed before its mhwd dependency
warning: dependency cycle detected:
warning: mhwd-db will be installed before its mhwd dependency

Packages (2) mhwd-0.6.5-29  mhwd-db-0.6.5-29

Total Download Size:   0,09 MiB
Total Installed Size:  0,23 MiB
Net Upgrade Size:      0,00 MiB

:: Proceed with installation? [Y/n] y
:: Retrieving packages...
 mhwd-0.6.5-29-x86_64                                                           82,5 KiB  1756 KiB/s 00:00 [###############################################################] 100%
 mhwd-db-0.6.5-29-any                                                           12,2 KiB  1740 KiB/s 00:00 [###############################################################] 100%
 Total (2/2)                                                                    94,7 KiB  1280 KiB/s 00:00 [###############################################################] 100%
(2/2) checking keys in keyring                                                                             [###############################################################] 100%
(2/2) checking package integrity                                                                           [###############################################################] 100%
(2/2) loading package files                                                                                [###############################################################] 100%
(2/2) checking for file conflicts                                                                          [###############################################################] 100%
(2/2) checking available disk space                                                                        [###############################################################] 100%
:: Running pre-transaction hooks...
(1/1) Creating Timeshift snapshot before upgrade...
==> skipping timeshift-autosnap due skipRsyncAutosnap in /etc/timeshift-autosnap.conf set to TRUE.
:: Processing package changes...
(1/2) upgrading mhwd-db                                                                                    [###############################################################] 100%
(2/2) upgrading mhwd                                                                                       [###############################################################] 100%
:: Running post-transaction hooks...
(1/3) Arming ConditionNeedsUpdate...
(2/3) Updating linux initcpios...
==> Building image from preset: /etc/mkinitcpio.d/linux66.preset: 'default'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-6.6-x86_64 -g /boot/initramfs-6.6-x86_64.img --microcode /boot/intel-ucode.img
==> Starting build: '6.6.10-1-MANJARO'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [autodetect]
  -> Running build hook: [modconf]
  -> Running build hook: [block]
==> WARNING: Possibly missing firmware for module: 'xhci_pci'
  -> Running build hook: [keyboard]
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [resume]
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating gzip-compressed initcpio image: '/boot/initramfs-6.6-x86_64.img'
==> Image generation successful
==> Building image from preset: /etc/mkinitcpio.d/linux66.preset: 'fallback'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-6.6-x86_64 -g /boot/initramfs-6.6-x86_64-fallback.img -S autodetect --microcode /boot/intel-ucode.img
==> Starting build: '6.6.10-1-MANJARO'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [modconf]
  -> Running build hook: [block]
==> WARNING: Possibly missing firmware for module: 'aic94xx'
==> WARNING: Possibly missing firmware for module: 'qed'
==> WARNING: Possibly missing firmware for module: 'qla2xxx'
==> WARNING: Possibly missing firmware for module: 'wd719x'
==> WARNING: Possibly missing firmware for module: 'qla1280'
==> WARNING: Possibly missing firmware for module: 'bfa'
==> WARNING: Possibly missing firmware for module: 'xhci_pci'
  -> Running build hook: [keyboard]
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [resume]
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating gzip-compressed initcpio image: '/boot/initramfs-6.6-x86_64-fallback.img'
==> Image generation successful
(3/3) Refreshing PackageKit...
[nls@lap ~]$ date -u
sam. 27 janv. 2024 11:21:40 UTC
[nls@lap ~]$ 

steam is working, intel cpu & gpu on my laptop.

Just disable integrated graphics in the BIOS. Where that option is depends on BIOS version;

Why should he update his Bios?

Im not sure if this related to AMD Mainboards, because AMD throwing infinity Bios updates like a rolling release… but for Intel Mainboards there is still the old rule… never change a running system and don’t update your Bios when you don’t have bugs.

The most issues are not fixed by a bios update… atleast this is my 30years old experience with PC’s.

Exactly… this is right. Anyways, AMD has alot issues build in with his stupid (since ryzen release) aegis updates or how they called.

Some few and special games run into issue sometimes and in this situation, from my weak AMD Ryzen Mainboard unterstanding (this is the reason, why i don’t buy AMD Mainboards and choose Intel, when possible), but you shouldn’t run into issue with all games… why i think it doesn’t make sense to update your Bios right now.

Hello,

I did try that, as it makes sense, but when doing so, I simply get no display output. Which is peculiar because I am plugged into my dGPU for graphics, and as far as I know, the iGPU cannot work its way through a discrete card to provide graphics. I cannot even view my BIOS/UEFI when I disabled the iGPU.

I did use this GPU in a Windows machine prior to switching to Linux, and it worked as expected. And in the handful of games that I can get working, performance is in line with performance expected from the card, as opposed to what the integrated graphics would be capable of.

Are you suggesting I do, or do not update the BIOS with the aegis (as you called it) updates? I am unsure if that’s something related to my issue, or more of a general statement. By-and-large I am unable to run games, but a few do work for me, which is the opposite situation of a few games being affected in the way you say.

There is likely another setting somewhere in the BIOS which says which is the Primary Video Adaptor - integrated graphics (IGD) or external graphics (PEG). If so set it to external, the default setting is probably integrated graphics.

If not then I have no idea.

Something is wonky …
Its already been acting like its stuck on the dGPU only.
But setting it to that in BIOS breaks things?

IDK.

The inxi is confusing as well …

Note the DRI indicates the amd/radeon

But then all the API stuff indicates the nvidia

Except vulkan which lists both, in their correct designations

And remember above - the default outputs used nvidia … while prime-run didnt work at all.
(though this may have changed after reinstalling the updated mhwd profile?)

PS and to the side …

As a general rule you should. Especially with newer hardware.
I didnt actually tell them to do it, or state it was related … just mentioned it was out of date.
Looking at the available updates … they improve performance, compatibility, and close security holes.
If you avoid them because of some vague sense of ‘bios update scary’ … then I guess thats your prerogative. :person_shrugging:

Im not sure if this a AMD iGPU thing, but with my Skylake i had disabled my iGPU on the first day and never experienced something like that.

I think this shouldn’t be normal, maybe its worth to investigate future… from my view point, you don’t need to have a iGPU Enabled when you have a good dGPU.

As i mentioned, i never have owned a Ryzen PC for this particular reason… always flashing your UEFI is a chore.

I don’t have first hand experience with Ryzen Hardware (besides my AMD Laptop)… which i had luckylie never updated yet, because tuxedo send it to me with a fresh bios and since its not the newest brand of technologie, i hope i don’t have too.

Anyways, since you can’t even deactivate your iGPU even when you had your Monitor connected with your nvidia 3080 and you see no signal, i think you may have a Bios bug there… and probably need to do your chore, since you choosed AMD :sweat_smile:

I also have to add, that Linux don’t like dual Monitor to much, i probably would try to connect just 1 monitor… but its just a blind guess. With only 1 monitor connected, i would probably also try to deactivate your iGPU again… also changing the display ports, when you have a blank screen.

I would also recommend to use a SWAP Partition, i monitored swap usage in Windows for years and i can tell Linux do it much better.

The swap partition/file will mainly not used as you have free RAM (in Linux), but for Windows the usage is crazy dumb.

I wouldn’t worried about the wear from your SSD in Linux because of swap, not with 32GByte RAM… but it can’t hurt.

Indeed it has! Running glxinfo tells me the renderer is the RTX 3080. That is with and without prime-run. Which means that the dGPU is being used by default as intended.

I can definitely try updating my BIOS, and possibly it is the magic fix I need. I can also add some swap space like @Kobold suggests and see if that helps with the tanking performance when launching a game.

I hope the issue isn’t having multiple monitors. I would rather not be able to play some games than be limited to a single monitor, to be honest.

Alright … I wonder ( if we may just have a functionally reverse-prime by accident ? )…
What about:

DRI_PRIME=0 glxinfo | grep 'renderer string'
DRI_PRIME=1 glxinfo | grep 'renderer string'

I went and updated my BIOS to the newest version.

Output for both was the GPU (OpenGL renderer string: NVIDIA GeForce RTX 3080/PCIe/SSE2). iGPU is still enabled in BIOS, because disabling it caused issues. There is also a setting there about detecting if the display output is on the dGPU and adjust accordingly, but that was off. So, not sure about that, but that’s not something I’m worried about messing with right now.

Definitely not a thing on earlier AMD Ryzen, I ran Manjaro on integrated 5700G graphics for many months when I first build my current PC because gfx cards were still insanely expensive at the time. When I did eventually get a discrete card there was no problem at all with disabling the igpu.

It makes no sense that disabling the igpu would prevent output from the discrete card, there must be some other BIOS setting that is causing that.

Like that… :thinking:

By the way (not related to the topic): How about searching for Pacnew files :see_no_evil:

I went to check my BIOS/UEFI settings, and here are the relevant options:

[Image embed was prohibited, but it was the UEFI menu showing display priority, with the options being Internal and External, where External is selected]

The display priority is set to External, as the alternative is Internal since I didn’t really want internal graphics. I assume that means integrated vs dedicated, but maybe it means like eGPUs like with thunderbolt connections and such, in which case it’s possible that is where the issue comes from.

[Image embed was prohibited, but it showed the menu at Advanced\AMD CBS\NBIO Common Options\GFX Configuration which has “Auto Detect Discrete Display” set to “Disabled” and the description of that option is “Detect discrete display to disable iGPU”. Beneath that option is 'iGPU Configuration" which is set to “UMA_AUTO” which is default. The other settings in the menu are “UMA Version” and “GPU Host Translation Guide” which are both set to “Auto”.]

For the “detect discrete display” setting, it is set to disabled, which is default, as how I read it was that having the feature enabled would allow the computer to disable the iGPU if it thinks the dGPU is supposed to be used. From how I understand it, leaving it as-is shouldn’t be causing any issues.

Back to the main issue at-hand, I realize I put my logs in the tmp directory, which has been cleared since making the log (because duh it’s temporary), and using “steam 2>&1 | tee /tmp/dumps” no longer creates the log I saw before, which means I probably forgot some setup I need to do in order to generate a log file that can be useful. Apologies to those on this thread for my being rather hapless.

Looking at logs, the following lines are the first to show up in red (there are yellow-colored) lines which don’t appear to be a problem

err:winediag:nodrv_CreateWindow Application tried to create a window, but no driver could be loaded.
err:winediag:nodrv_CreateWindow L"The explorer process failed to start."

That appears in all the games that don’t launch for me. The following line is “Setting breakpad minidump AppID = _____”, but looking it up that doesn’t seem to be an indicator of a problem.
For some games, the same error appears again shortly after, and then the game tries to shutdown. But other games go for a while longer until another red-text/error message. One of the games I am looking at the logs for (Palworld), gets a few error messages sprinkled for a while until eventually giving:

err:vulkan:init_vulkan Failed to load Wine graphics driver supporting Vulkan.
err:vulkan:init_vulkan Failed to load Wine graphics driver supporting Vulkan.

With both of those errors one after the other, and then the game terminates.

Not sure if this will help any, just throwing some darts to see what sticks, but I saw a thread suggesting to run “DISPLAY=:0 winecfg” which for me gave the following output, with the whole computer becoming extremely sluggish like when I launch a game (even ones that work), before eventually winecfg launches and works fine after the delay

007c:fixme:hid:handle_IRP_MN_QUERY_ID Unhandled type 00000005
007c:fixme:hid:handle_IRP_MN_QUERY_ID Unhandled type 00000005
007c:fixme:hid:handle_IRP_MN_QUERY_ID Unhandled type 00000005
007c:fixme:hid:handle_IRP_MN_QUERY_ID Unhandled type 00000005
008c:err:winediag:is_broken_driver Broken NVIDIA RandR detected, falling back to RandR 1.0. Please consider using the Nouveau driver instead.
00e4:err:winediag:is_broken_driver Broken NVIDIA RandR detected, falling back to RandR 1.0. Please consider using the Nouveau driver instead.
008c:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
008c:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
008c:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
00e4:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
00e4:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
00e4:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
008c:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
008c:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
008c:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
008c:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
008c:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
008c:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
0034:err:winediag:is_broken_driver Broken NVIDIA RandR detected, falling back to RandR 1.0. Please consider using the Nouveau driver instead.
00dc:err:winediag:is_broken_driver Broken NVIDIA RandR detected, falling back to RandR 1.0. Please consider using the Nouveau driver instead.
0124:err:winediag:is_broken_driver Broken NVIDIA RandR detected, falling back to RandR 1.0. Please consider using the Nouveau driver instead.
00e4:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
00e4:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.
00e4:fixme:xrandr:xrandr10_get_current_mode Non-primary adapters are unsupported.

Not sure how much of that is actually problematic or just the computer whining about something not being ideal that is inconsequential. If nothing else, maybe someone else having a similar issue stumbles upon this from some specific string searches and get some helpful info :slight_smile:

I tried some things that seem to have worked for the games I tested that didn’t work before. It seems that whatever installing I did for graphics drivers before through mhwd and the normal package manager didn’t quite work. Also worth noting that I still have nvidia-prime installed, but it seems fine? Using lspci -k | grep -EA3 'VGA|3D|Display' shows both my GPU and my CPU with integrated graphics, but really all I need working is the GPU. So; resolved!

What I did that worked:
1.) pacman -Rns winetricks
2.) pacman -Rns wine
3.) I manually deleted the .wine config folder in my home directory
4.) I also searched in “Add/Remove Packages”
4.) I then followed GitHub - korvahannu/arch-nvidia-drivers-installation-guide: A repository that holds a simple guide on how you can install and configure proprietary NVIDIA drivers for Arch Linux.
Note that some of the steps in that link weren’t exactly the same; for instance I was using the base “nvidia” drivers, and one of the steps at the end mentions changing a hook thing to the example given, but for me it already had the correct driver listed for the hook, so I didn’t change it. Also I didn’t have a /etc/pacman.d/hooks/ directory, so I manually made that before putting the nvidia.hook in it.

This topic was automatically closed 36 hours after the last reply. New replies are no longer allowed.