Hi @jpleb, please help test with our latest driver release - 560.31.02. Thank you
1 Like
@abchauhan was this info enough to file an internal ticket about HDR?
Hi @lord, Yes, we are looking for suitable hardware for the HDR issues. I will reach out if I need more information. Thanks
1 Like
Hi all, listing bugs from this feedback & discussion :-
- 4764568 - Minecraft crashes when entering full screen
→ Under investigation
- 4767941 - All Games run using Wine or proton seems to be crashing
X Error of failed request: BadValue (integer parameter out of range for operation)
Major opcode of failed request: 146 ()
Minor opcode of failed request: 5
Value in failed request: 0x1c003ba
Serial number of failed request: 3986
Current serial number in output stream: 4003
0210:fixme:kernelbase:AppPolicyGetProcessTerminationMethod FFFFFFFFFFFFFFFA, 000000000011FE80
→ Fixed in 560.31.02. Closing this bug.
- 4769138 - Unreal Engine crashing on launch
→ Fixed in 560.31.02. Closing this bug.
- 4769163 - Screen freeze observed with directly scanned out windows on a multi-monitor configuration
Jul 26 15:52:35 beefiercake kernel: [drm:nv_drm_atomic_apply_modeset_config.isra.0 [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to initialize semaphore for plane fence
Jul 26 15:52:35 beefiercake kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to apply atomic modeset. Error code: -11
Jul 26 15:53:00 beefiercake kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 2
Jul 26 15:53:03 beefiercake kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 2
→ Fixed. The fix will be available in our upcoming driver releases.
- 4545537 - Multi-monitor VRR support on Wayland with Explicit sync
→ Under investigation.
- 4757907 - Bug :nvidia-drivers-555.58 DVI output not working ->black screen
→ Fixed. The fix will be available in our upcoming driver releases. Please follow-up on - Bug :nvidia-drivers-555.58 DVI output not working ->black screen
- 4717891 - On 555, Unreal engine crashes after running for some time
CommonUnixCrashHandler: Signal=11
[775]LogCore: === Critical error: ===
Unhandled Exception: SIGSEGV: invalid attempt to read memory at address 0x0000000000000000
[775]LogCore: === Critical error: ===
Unhandled Exception: SIGSEGV: invalid attempt to read memory at address 0x0000000000000000
[775]LogCore: Fatal error!
0x00007e78fd4086b7 libnvidia-glcore.so.555.52.04!UnknownFunction(0xa086b7)
0x00007e78fd7e7079 libnvidia-glcore.so.555.52.04!UnknownFunction(0xde7078)
0x00007e78fd7fe780 libnvidia-glcore.so.555.52.04!UnknownFunction(0xdfe77f)
0x00007e78fd73aede libnvidia-glcore.so.555.52.04!UnknownFunction(0xd3aedd)
0x00007e78fc799fdf libUnrealEditor-VulkanRHI.so!VulkanRHI::FFenceManager::WaitForFence(VulkanRHI::FFence*, unsigned long long) [/home/4717891/UnrealEngine-5.3.0-release/Engine/Source/./Runtime/VulkanRHI/Private/VulkanMemory.cpp:4660]
0x00007e78fc73fe7d libUnrealEditor-VulkanRHI.so!FVulkanCommandBufferManager::WaitForCmdBuffer(FVulkanCmdBuffer*, float) [/home/4717891/UnrealEngine-5.3.0-release/Engine/Source/./Runtime/VulkanRHI/Private/VulkanCommandBuffer.cpp:502]
0x00007e78fc85a839 libUnrealEditor-VulkanRHI.so!FVulkanViewport::WaitForFrameEventCompletion()
→ Under investigation.
- 4770124 - NVIDIA Vulkan ICD wakes up dgpu on initialization and exit
→ Under investigation. Please follow-up on - [550.67] Nvidia Vulkan ICD wakes up dgpu on initialization and exit
- 200504689 - Remove “P2 forced” state from drivers
→ Under investigation. Please follow-up on - Remove "P2 forced" state from drivers
- Spotify fails to create a dri2 screen
- Various applications - libEGL warning: egl: failed to create dri2 screen
- Electron applications appear blank
- Screen capture appears blank
- Missing characters on Firefox text, HW acceleration failing on Firefox
→ Fixed on 560.31.02/ egl-wayland 1.1.14
- 4797139 - Overriding EDID makes VRR stop working under WAYLAND → ( “vrr_capable” (immutable): range [0, 1] = 0 ) with driver 560.31.02
→ Under investigation. We can reproduce this on our systems.
- 4806756 - Xwayland VRAM usage is still excessive when resizing X11 apps under wayland with driver 560.28.03
→ Under investigation.
- 4821165 - Screen freeze observed when Night Mode is toggled on Ubuntu 24.04 PRIME notebooks with an external monitor
2024-08-20T19:35:18.662981+02:00 rog kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
2024-08-20T19:35:21.798987+02:00 rog kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
→ Under investigation. Please follow-up on - Screen freezes at random intervals with RTX 4060 Max-Q / Mobile, multiple driver versions tested
- 4821290 - Colours appear washed out on multiple displays - Deco Gear 40, LG OLED C1, Alienware AW2725DF, Alienware AW3423DWF and Samsung Odyssey G8 OLED, when HDR is enabled
→ Checking if we can reproduce this on our systems.
amrits and I are checking if we can reproduce the HDR issues, XWayland memory issues.
We are reviewing this feedback page for other issues. We will file bugs and share bug IDs for tracking.
Thank you
11 Likes
560.31.02 is what I have been running and what I was running when I reported this. GTX 1070, using the nvidia-beta-dkms AUR package.
I should add that this issue also causes OBS to freeze and crash sometimes. The entire desktop will lock up for a few seconds, then OBS will close.
If you need some help to debug HDR issue, please write.
Not really 560 specific but I thought I’d shine a light on this, with Above 4G decoding enabled in BIOS, Control doesn’t utilize the full GPU.
If I disable Above 4G Decoding in BIOS, it utilize the full GPU in the game (just disabling Rebar alone does not help).
See screenshots.
4G enabled (73% GPU usage):
4G disabled (100% GPU usage and there’s roughly a 15FPS gain):
I’m not sure how many games are affected by this but it’s not a general solution just to disable Above 4G decoding since it usually makes things worse.
I have no real idea if it’s the driver, or possibly vkd3d-proton that’s causing this.
(Windows is not affected)
I can also help debugging memory leak.
tldr; to reproduce: resizing any window, in example terminal window (few times) with nvtop will show increase huge VRAM consumption (from 0.7GB to 2.5GB) and is never released. Gnome DE.
1 Like
nvidia-bug-report.log.gz (449.0 KB)
I have been encountering a bug in starfield since the game’s release on Arch Linux, since driver 535 if I remember correctly, where my RTX 4090 TUF GAMING OC maxes out at about 196W of power (measured through mangohud) and a maximum of 60fps at a varying resolutions, from 4k @ 120hz to 5120x1440p @ 240hz. sadly the game does not seem to EVER go above 60fps regardless of video settings, at first I believed it to be a bug with the game but on windows the game performs perfectly, I get nearly triple the performance at the same video settings without DLSS FG.
Which leads me to believe it to be some form of bug with either the game itself or the drives since the release of the game, sadly I cannot find anyone with the same issue, and the game is unmodded, as vanilla as it can be, using a fresh installation of CachyOS, I have tried proton Experimental, 9, 8, 7 and GE since the game’s first release to no avail.
the game is running on a SP XS70 4TB, R9 7900x with 64GB 6000MT/s RAM, and this is the only game that experiences such behavior to my knowledge. I would appreciate assistance in this matter as I believe only Nvidia or Bethesda can resolve this issue now.
Best Regards.
Yes, all info that Nvidia could possibly need is here: Xwayland VRAM usage is still excessive when resizing X11 apps under wayland. · Issue #126 · NVIDIA/egl-wayland · GitHub
What I find odd in your case is that it stops at 2,5GB, for me it stops allocating at 1235MB but like you said, it’s never released.
It depends on resolution i think. More i resize the window more vram is consumed.
1 Like
An issue about nv_queue crashes is found from the 555 release, and it’s still not fixed,both in close and open source drivers, and maybe it’ll cause random kernel panics.
You can find related information at here:
and
To reproduce this problem, you can try to run rog control center on a asus laptop with both amd igpu and nvidia dgpu.You’ll see this log in dmesg after you starts this app.
(rog-control-center is not related with this issue, as some other people also experienced the same issue without it, it’s just the fastest way to reproduce the problem :-)
I can also reproduce the memory leak by resizing it like @ppogorze mentioned, for me it also stops at around 1000MiB
Any tracking number for the g-sync / vrr high power usage?
Here is the screenshot with VRR disabled in Hyprland settings. nvtop shows usage of around 12W and the GPU mem clock being 405Mhz.
Now this screenshot is after I turn on VRR / GSync. You can now see the GPU using over 30W idling and also it’s memory clock being maxxed out to 8000Mhz.
The game is the unreleased Deadlock by Valve. Since it’s not well optimised yet, it’s a good benchmark for me since I can fill the VRAM in minutes rather than hours playing it (this issue does happen with other games, just takes a lot longer). I ran it with DX11 and upped the FSR2 from balanced to quality to accelerate the issue.
I’ve generated the report right after the VRAM hit 10GB and I had a stutter. I was averaging 100 FPS when VRAM was below 9.6GB, then it started dipping towards 70 FPS and as soon as it hit 10GB and the stutter stopped it was back to 100 FPS again.
nvidia-bug-report.log.gz (2.4 MB)
On a side note, I’ve decided to plug my second monitor on my motherboard and enable my iGPU (I have a Ryzen 7 7800X3D) and by doing so, VRR is working as intended on my main monitor without having to disable the secondary.
I have a Odyssey Neo G7 which doesn’t have any way to check the current refresh rate on the OSD, but the usual stutters without VRR enabled are quite noticeable when a game is going from 70 to 100 FPS this much, however I can’t notice any stutter with VRR on (as is intended), and this is the case now. I know this multi monitor VRR issue is being tracked internally, but just wanted to add some information that could help.
EDIT: I forgot to mention that plugging my secondary monitor on my iGPU also “fixes” the high power draw on idle on my card. With both monitors plugged to my 3080, it never goes below 100W no matter what, but when I plug my secondary on my iGPU, my 3080 goes down to 40W when idle.
I’m afraid it’s still failing to boot with Vulkan option enabled. Whereas others on AMD gpus are successful in launching and playing the game.
Edit: here is the logs just after the game fails to launch -
nvidia-bug-report.log.gz (473.7 KB)
Issue with the game THE FINALS not utilizing GPU fully.
Core clocks are very unstable and sometimes locks my GPU to 1200mhz in that game. Should be 1845mhz?
Issue in both Wayland and X11.
CPU: AMD Ryzen 7 5800X 8-Core Processor
Kernel 6.10.4-1-cachyos-sched-ext-lto
Driver: NVIDIA GeForce RTX 2080/PCIe/SSE2
Driver: NVIDIA 560.31.02
nvidia-bug-report.log.gz (452.2 KB)
I also play finals and get much worse performance under linux with my RTX 3070FE. Hopefully nvidia team can investigate.
This has nothing to do with 555 or 560 specifically.
The game causes clocks to fluctuate wildly for no reason, maybe because of bad Reflex integration.
i get poor fps and utilization with and without reflex.