protonscr

Final Fantasy VII Rebirth

protonopen appid 2909400Game compatibility - UnofficialNVIDIA driversXAudio2Mesa driversAMD RADV
ValveSoftware/Proton#8408 · opened 2025-01-23 by hjpaul7 · updated 2026-08-14 · 342 comments · github · game page · search this game
12 matching comments, n / p to jump
Hhjpaul7 2025-01-23 github

Compatibility Report

  • Name of the game with compatibility issues: Final Fantasy VII Rebirth
  • Steam AppID of the game: 2909400

System Information

  • GPU: Nvidia 4090
  • Video driver version: 565.77
  • Kernel version: 6.12.10
  • Link to full system information report: Gist here
  • Proton version: Experimental

I confirm:

  • [x] that I haven't found an existing compatibility report for this game.
  • [x] that I have checked whether there are updates for my system available.

Proton Log: steam-2909400.log

Symptoms

Missing textures once you're able to play the game (after cutscene). I'm assuming there's supposed to be a UI tutorial prompt being displayed but cannot see it as you're unable to move the character.

Reproduction

  • Launch game
  • Get through opening cutscenes till you're able to play the game
  • Textures missing (screenshot below)

Image

UUklosk 2025-01-23 github

Same here with an RTX 3080.

HHansKristian-Work 2025-01-23 github

The RTX issues are known. It's a driver bug and it has been reported to NV with bug number 5059946.

I've spent a lot of time trying to figure out a reasonable workaround, but to no avail.

Hhjpaul7 2025-01-23 github

The RTX issues are known. It's a driver bug and it has been reported to NV with bug number 5059946.

I've spent a lot of time trying to figure out a reasonable workaround, but to no avail.

@HansKristian-Work Thanks. Didn't Final Fantasy 16 have the same issue at launch? Curious if that was fixed via a driver release.

DDinjoralo 2025-01-23 github

The RTX issues are known. It's a driver bug and it has been reported to NV with bug number 5059946.

I've spent a lot of time trying to figure out a reasonable workaround, but to no avail.

Thank you for this information. I'm sorry to be a bother, but would you be aware of any timeframe for when the next Nvidia driver may be released? Whether it includes the fix for this specific game or not, there's a lot of problems it's supposed to address that I and many others are anxious for.

Sseidtgeist 2025-01-23 github

I'm getting this whenever I launch:

DX12 is not supported on your system. Try running without the -dx12 or -d3d12 command line argument.

Running a 1080 Ti on 565.77.

Edit:

Setting this as launch options allowed it to launch, now compiling shaders:

VKD3D_CONFIG=dxr VKD3D_FEATURE_LEVEL=12_2 VKD3D_SHADER_MODEL=6_6 %command%

Edit 2: Tried some different combinations. This error always appears after the end of shader compilation:

Image

Edit 3: See @shigutso's comment, thank you for the detailed information.

DDistantThunder 2025-01-23 github

Note: Gamescope (HDR)

When running the game with gamescope for HDR, it initialy detects and is stuck at 720p for some reason no matter what one sets as Gamescope option for resolution.

Literally juggling haphazardly between Fullscreen/Windowed allowed me to gain actual targeted resolution. This should be immediately clear as all blurriness is gone from the menu itself.

Rroccodev 2025-01-23 github

A quick look from the AMD side (5700X, 6700XT):

  • The game seems to have a native shader compilation step on first boot, though for some reason it crashed on completion. When you boot the game again it skips this step so you can play just fine.
  • The game can't find save data from Remake, likely because they are in different prefixes. As with Metaphor: ReFantazio and its demo, the solution is probably to copy the Remake save data into Rebirth's prefix.
  • The game uses full screen by default, though resolution is reported as 1920x1080 (on a 4K screen). The game is still outputting 4K though (also confirmed by frame rate). This can be fixed by selecting Windowed, changing the resolution to 4K, then selecting Full Screen again.
  • Sometimes, there are noticeable LOD pop-ins at the start of cutscenes or when moving the camera quickly.
  • Perhaps related to the previous point, frame rate is quite unstable when moving the camera (~55-60 dropping to ~47 FPS). Admittedly I'm running quite ambitious settings for the hardware (max settings except Low Shadows, and 66-100% resolution scaling), and I think it might be related to the dynamic resolution, because pinning it to 50-50% gets rid of the stutters (but it might also be due to the lower resolution)
  • When I was fiddling with the graphics settings without any manual saves yet, I noticed that the game would always start with Japanese voices, even though I was making sure that it was set to English in the menu. Other settings would apply just fine. Restarting the game fixed this, I wonder if it is also an issue on Windows.
EEoinHealy0 2025-01-23 github

A quick look from the AMD side (5700X, 6700XT):
The game can't find save data from Remake, likely because they are in different prefixes. As with Metaphor: ReFantazio and its demo, the solution is probably to copy the Remake save data into Rebirth's prefix.

You're correct on the save profile, that worked for me earlier today

ZZarkorix 2025-01-23 github

The RTX issues are known. It's a driver bug and it has been reported to NV with bug number 5059946.
I've spent a lot of time trying to figure out a reasonable workaround, but to no avail.

Is there a link to the bug report so we can track the progress? Also I'm experiencing the exact same issue, RTX 3060Ti + 565.77.

DDuplicantWood 2025-01-23 github

The RTX issues are known. It's a driver bug and it has been reported to NV with bug number 5059946.
I've spent a lot of time trying to figure out a reasonable workaround, but to no avail.

Thank you for this information. I'm sorry to be a bother, but would you be aware of any timeframe for when the next Nvidia driver may be released? Whether it includes the fix for this specific game or not, there's a lot of problems it's supposed to address that I and many others are anxious for.

According to phoronix the next driver is out on the 30th.

https://www.phoronix.com/news/GeForce-RTX-5090-Linux-Driver

Oofcaah 2025-01-23 github

The RTX issues are known. It's a driver bug and it has been reported to NV with bug number 5059946.
I've spent a lot of time trying to figure out a reasonable workaround, but to no avail.

Is there a link to the bug report so we can track the progress? Also I'm experiencing the exact same issue, RTX 3060Ti + 565.77.

Same deal with 3090. Game's pretty much unplayable. Hopefully some workaround surfaces, or it'll probably be a while before we get drivers with a fix. There's a phoronix article stating it'll be in a week, but will it fix this problem?

As to the bug tracker, according to nvidia dev portal:

  1. URL should be https://developer.nvidia.com/bugs/5059946
  2. however it can be only viewed by submitter.
Eelken 2025-01-23 github

Another AMD report (7700X, 7800XT) on Bazzite

Currently unable to watch the first cutscene or get through the first area without KWin crashing. Taking a while to understand why, because it's right at the end so every change takes 2-3m even with fast-forward.

Seems to be playtime dependent, after 5 minutes of usage like this (note the clock speed too...) it just bombs.

Image

Confirm the copy save trick works for getting the bonuses so that's a quick win.

Ddmsimard 2025-01-23 github

I am not sure how to distinguish whether this is an issue with the game, proton, the amdgpu driver or something else but I am sharing just in case.

Running a Ryzen 5800X with a 6800XT on an up-to-date Fedora 41 with Kernel 6.12.9-200.fc41.x86_64.
(Edit: glxinfo Device: AMD Radeon RX 6800 XT (radeonsi, navi21, LLVM 19.1.5, DRM 3.59, 6.12.9-200.fc41.x86_64) (0x73bf))

The game launches fine and I managed to play until Nibelheim where I get repeated crashes.
The gnome desktop session crashes and it sends me back to login.

On another note, the indoor lighting seems WAY too bright.

I tried proton experimental as well as the latest GE proton (9-23) to no avail.

In dmesg I see the following:

[Thu Jan 23 13:36:21 2025] amdgpu 0000:09:00.0: amdgpu: Dumping IP State
[Thu Jan 23 13:36:21 2025] amdgpu 0000:09:00.0: amdgpu: Dumping IP State Completed
[Thu Jan 23 13:36:21 2025] amdgpu 0000:09:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=145837, emitted seq=145839
[Thu Jan 23 13:36:21 2025] amdgpu 0000:09:00.0: amdgpu: Process information: process ff7rebirth_.exe pid 9031 thread vkd3d_queue pid 9174
[Thu Jan 23 13:36:21 2025] amdgpu 0000:09:00.0: amdgpu: GPU reset begin!
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: MODE1 reset
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: GPU mode1 reset
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: GPU smu mode1 reset
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: GPU reset succeeded, trying to resume
[Thu Jan 23 13:36:22 2025] [drm] PCIE GART of 512M enabled (table at 0x0000008000F00000).
[Thu Jan 23 13:36:22 2025] [drm] VRAM is lost due to GPU reset!
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: PSP is resuming...
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: reserve 0xa00000 from 0x83fd000000 for PSP TMR
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: SMU is resuming...
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: smu driver if version = 0x00000040, smu fw if version = 0x00000041, smu fw program = 0, version = 0x003a5a00 (58.90.0)
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: SMU driver if version not matched
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: use vbios provided pptable
[Thu Jan 23 13:36:22 2025] amdgpu 0000:09:00.0: amdgpu: SMU is resumed successfully!
[Thu Jan 23 13:36:22 2025] [drm] kiq ring mec 2 pipe 1 q 0
[Thu Jan 23 13:36:22 2025] [drm] DMUB hardware initialized: version=0x02020020
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring gfx_0.1.0 uses VM inv eng 1 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 4 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 5 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring kiq_0.2.1.0 uses VM inv eng 12 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring sdma0 uses VM inv eng 13 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring sdma1 uses VM inv eng 14 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring sdma2 uses VM inv eng 15 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring sdma3 uses VM inv eng 16 on hub 0
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring vcn_dec_0 uses VM inv eng 0 on hub 8
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc_0.0 uses VM inv eng 1 on hub 8
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc_0.1 uses VM inv eng 4 on hub 8
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring vcn_dec_1 uses VM inv eng 5 on hub 8
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc_1.0 uses VM inv eng 6 on hub 8
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc_1.1 uses VM inv eng 7 on hub 8
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: ring jpeg_dec uses VM inv eng 8 on hub 8
[Thu Jan 23 13:36:23 2025] amdgpu 0000:09:00.0: amdgpu: GPU reset(2) succeeded!
[Thu Jan 23 13:36:23 2025] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[Thu Jan 23 13:36:23 2025] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
AAshtenThePlatypus 2025-01-23 github

Running a Ryzen 5800X with a 6800XT on an up-to-date Fedora 41 with Kernel 6.12.9-200.fc41.x86_64. (Edit: glxinfo Device: AMD Radeon RX 6800 XT (radeonsi, navi21, LLVM 19.1.5, DRM 3.59, 6.12.9-200.fc41.x86_64) (0x73bf))

The game launches fine and I managed to play until Nibelheim where I get repeated crashes. The gnome desktop session crashes and it sends me back to login.

Can confirm, the errors you pasted from your logs also show up on AMD 6900XT, Ubuntu 24.04, Kernel 6.8.0-51-generic.

[  301.193915] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, signaled seq=141858, emitted seq=141860
[  301.194494] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process ff7rebirth_.exe pid 26475 thread vkd3d_queue pid 26580
[  301.195042] amdgpu 0000:09:00.0: amdgpu: GPU reset begin!
[  301.600324] amdgpu 0000:09:00.0: amdgpu: MODE1 reset
[  301.600331] amdgpu 0000:09:00.0: amdgpu: GPU mode1 reset
[  301.600401] amdgpu 0000:09:00.0: amdgpu: GPU smu mode1 reset
[  302.104043] amdgpu 0000:09:00.0: amdgpu: GPU reset succeeded, trying to resume
[  302.107303] [drm] PCIE GART of 512M enabled (table at 0x0000008002000000).
[  302.107403] [drm] VRAM is lost due to GPU reset!
[  302.107406] [drm] PSP is resuming...
[  302.186651] [drm] reserve 0xa00000 from 0x83fd000000 for PSP TMR
[  302.328245] amdgpu 0000:09:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
[  302.328257] amdgpu 0000:09:00.0: amdgpu: SMU is resuming...
[  302.328263] amdgpu 0000:09:00.0: amdgpu: smu driver if version = 0x00000040, smu fw if version = 0x00000041, smu fw program = 0, version = 0x003a5900 (58.89.0)
[  302.328270] amdgpu 0000:09:00.0: amdgpu: SMU driver if version not matched
[  302.328305] amdgpu 0000:09:00.0: amdgpu: use vbios provided pptable
[  302.407854] amdgpu 0000:09:00.0: amdgpu: SMU is resumed successfully!
[  302.409133] [drm] DMUB hardware initialized: version=0x02020020
[  302.657197] [drm] kiq ring mec 2 pipe 1 q 0
[  302.663034] [drm] VCN decode and encode initialized successfully(under DPG Mode).
[  302.663288] [drm] JPEG decode initialized successfully.
[  302.663304] amdgpu 0000:09:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[  302.663307] amdgpu 0000:09:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[  302.663309] amdgpu 0000:09:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[  302.663311] amdgpu 0000:09:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 5 on hub 0
[  302.663313] amdgpu 0000:09:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 6 on hub 0
[  302.663315] amdgpu 0000:09:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0
[  302.663317] amdgpu 0000:09:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0
[  302.663318] amdgpu 0000:09:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 9 on hub 0
[  302.663320] amdgpu 0000:09:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 10 on hub 0
[  302.663322] amdgpu 0000:09:00.0: amdgpu: ring kiq_0.2.1.0 uses VM inv eng 11 on hub 0
[  302.663324] amdgpu 0000:09:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
[  302.663326] amdgpu 0000:09:00.0: amdgpu: ring sdma1 uses VM inv eng 13 on hub 0
[  302.663329] amdgpu 0000:09:00.0: amdgpu: ring sdma2 uses VM inv eng 14 on hub 0
[  302.663330] amdgpu 0000:09:00.0: amdgpu: ring sdma3 uses VM inv eng 15 on hub 0
[  302.663332] amdgpu 0000:09:00.0: amdgpu: ring vcn_dec_0 uses VM inv eng 0 on hub 8
[  302.663334] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc_0.0 uses VM inv eng 1 on hub 8
[  302.663336] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc_0.1 uses VM inv eng 4 on hub 8
[  302.663338] amdgpu 0000:09:00.0: amdgpu: ring vcn_dec_1 uses VM inv eng 5 on hub 8
[  302.663340] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc_1.0 uses VM inv eng 6 on hub 8
[  302.663342] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc_1.1 uses VM inv eng 7 on hub 8
[  302.663344] amdgpu 0000:09:00.0: amdgpu: ring jpeg_dec uses VM inv eng 8 on hub 8
[  302.672107] amdgpu 0000:09:00.0: amdgpu: recover vram bo from shadow start
[  302.781922] amdgpu 0000:09:00.0: amdgpu: recover vram bo from shadow failed, r is -110, tmo is 0
[  302.781962] amdgpu 0000:09:00.0: amdgpu: GPU reset(2) failed
[  302.796148] amdgpu 0000:09:00.0: amdgpu: GPU reset end with ret = -5
[  302.796154] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* GPU Recovery Failed: -5
SShin-Aska 2025-01-23 github

VKD3D_CONFIG=dxr VKD3D_FEATURE_LEVEL=12_2 VKD3D_SHADER_MODEL=6_6 %command%

This one works on my GTX 1660S and gets into the game. Now I get the same issue as @hjpaul7 reported (textures missing)

Rroccodev 2025-01-23 github

@dmsimard @AshtenThePlatypus perhaps you should file a Mesa bug, I've had similar occurrences in Indiana Jones The Great Circle, apparently it's due to the game trying to compile an invalid shader or something.
But searches online also led me to think it was a voltage issue of sorts, after all it was a very demanding title (especially on the VRAM side). Are your cards factory overclocked? Personally I haven't had issues in Nibelheim (I stopped at the end of the reactor section, so maybe it's after that), but I also wasn't running it at max resolution (66% scaling)

NNullCGT 2025-01-23 github

Same issue as @hjpaul7 with an RTX4060. Like others, I am unfortunately unable to find a workaround. At the very least, we're not experiencing this issue alone.

Interestingly enough, I was able to move Zach around. I was interested in seeing if the issue persisted past this scene (I see no reason why it wouldn't), but unfortunately the invisible geometry proved too tricky to navigate.

Jjorbuedo 2025-01-24 github

4080 Super having the same missing terrain. I'm unfamiliar with how issue threads work here, should AMD and Nvidia issues be mixed on the same conversation? It's clearly different issues.
I'll advance to a different area with the Steam Deck and report whether it works then, maybe it's just the first area's texture.

Oofcaah 2025-01-24 github

should AMD and Nvidia issues be mixed on the same conversation?

They shouldn't, but welcome to the internet ;)

Ddabdude420 2025-01-24 github

im having the same problem with my nvidia 4080

Ddabdude420 2025-01-24 github

dose anyone now what the eta on fixing the problem mite be thank u

Jjorbuedo 2025-01-24 github

I'll advance to a different area with the Steam Deck and report whether it works then, maybe it's just the first area's texture.

No use, Nibelheim has the same issue. I'll have to install a Windows ssd for now 😿

DDistantThunder 2025-01-24 github
  • 9700x, 7900XTX, Mesa 24.3.3-2
  • Plasma 6.2.3-1
  • Arch Linux 6.12.7-zen1-1-zen

I did crash once at the beginning, playing at 4k@60 mostly, TSRU 100-66%. High. At chapter 2 currently.

Aaarron-lee 2025-01-24 github

Hardware: AMD 6700XT, 5600X
Distro: Bazzite

In game lighting seems to be a bit off, scenes will go from unnaturally dark to unnaturally bright, etc.

Also, I had hard crashes happening as soon as I would load a save file, but the crashing stopped after enabling V-sync both in-game and setting an fps via gamescope in gamescope-session. Testing more to see if this will crash at all.

Tteknolog1k 2025-01-24 github

In game lighting seems to be a bit off, scenes will go from unnaturally dark to unnaturally bright, etc.

Does the change in brightness occur consistently as you enter or leave a dark or bright area (ex: entering/exiting a building)? The game does do some weird brightness adjustments to mimic your eyes adjusting to a brighter or darker environment, but it's that way on PS5 as well, and seems like expected behavior.

Aaarron-lee 2025-01-24 github

Does the change in brightness occur consistently as you enter or leave a dark or bright area (ex: entering/exiting a building)? The game does do some weird brightness adjustments to mimic your eyes adjusting to a brighter or darker environment, but it's that way on PS5 as well, and seems like expected behavior.

Ah okay, so then it is expected behavior, good to know.

Rremz92 2025-01-24 github

endeavouros 5700x3d 7900xt

been getting some pretty hardcore crashes (whole system). seems slightly better under gamescope, but still happening. hope my gpu isn't dying...

Ggrahfmusic 2025-01-24 github

Like Ratchet and Clank FF7 Rebirth uses Direct Storage according to pcgamingwiki. I remember with Ratchet removing the direct storage dlls improved performance significantly when I first used it in Proton eons ago. Not sure if this is the case still or not.

Because of the use of DirectStorage 1.2 API, a M.2 NVMe SSD is generally recommended over a SATA SSD.
https://www.pcgamingwiki.com/wiki/Final_Fantasy_VII_Rebirth

Wwildan-zog 2025-01-24 github

I hope this issue fixed asap :')

Sshigutso 2025-01-24 github

I'm getting this whenever I launch:

DX12 is not supported on your system. Try running without the -dx12 or -d3d12 command line argument.

Running a 1080 Ti on 565.77.
Edit:
Setting this as launch options allowed it to launch, now compiling shaders:

VKD3D_CONFIG=dxr VKD3D_FEATURE_LEVEL=12_2 VKD3D_SHADER_MODEL=6_6 %command%

According to PCGamingWiki, the game requires DX12.2, not DX12 to work. It doesn't support Vulkan nor DX11.
My GPU, a RX 5500 XT, has support for up to DX12.1, not DX12.2, so I get the same error above when trying to launch the game.

The game's store page says DX12 "Ultimate", and now I realized that's actually DX12.2

For NVIDIA, GTX 20 series support DX12 Ultimate.
For AMD, 6000 series support it.

Hopefully there's a workaround to allow older GPUs to run the game. If the Steam Deck can run it, my RX 5500 XT can run it too. Requiring DX12.2 seems unnecessary...

Ssteckums 2025-01-24 github

It does not look like the 570 driver fixes the problem.

Image

Zzr00da 2025-01-24 github

In game lighting seems to be a bit off, scenes will go from unnaturally dark to unnaturally bright, etc.

Does the change in brightness occur consistently as you enter or leave a dark or bright area (ex: entering/exiting a building)? The game does do some weird brightness adjustments to mimic your eyes adjusting to a brighter or darker environment, but it's that way on PS5 as well, and seems like expected behavior.

Highly doubt this is the intentional look of the effect (7900 GRE)

Image

Ddmsimard 2025-01-24 github

Highly doubt this is the intentional look of the effect (7900 GRE)

Ryzen 5800X with a 6800XT on an up-to-date Fedora 41 with Kernel 6.12.9-200.fc41.x86_64.

Like you I am seeing the same thing indoors in Nibelheim and then shortly after that:

Image

00xmex 2025-01-24 github

It does not look like the 570 driver fixes the problem.

570 wasn't released officially, yet. That "leaked" release might be incomplete. Especially considering, that there are works on FF7R-2 currently going on. Wait for the official release, then test again.

EEoinHealy0 2025-01-24 github

You can also see that brightness issue happening on the Nvidia cards if you try play "guess where the ground is"

HHansKristian-Work 2025-01-24 github

Make sure you're using Proton Experimental, or those whiteout glitches are expected.

CCha1N1 2025-01-24 github

I use proton experimental bleeding edge and i still see white flashbangs full screen on mesa RADV drivers doesnt seem to happen on AMD pro drivers. (RX 6700 XT)

Oofcaah 2025-01-24 github

I've spent a lot of time trying to figure out a reasonable workaround, but to no avail.

Could you spare a little more detail? We can't read the bug report. There are numerous "mods" that replace/inject various DLLs, which apparently make the game workable on GTX 16XX series GPUs, and replace DLSS shipped with the game.

One of random URLs:
https://www.reddit.com/r/FFVIIRemake/comments/1i8mf4n/amdintel_gpus_you_can_mod_fsr_and_xess_into/

Do you believe such a "mod" could help work around the core issue we are suffering from, @HansKristian-Work ?

(to be precise: I'm not asking if replacing DLSS will fix this, but if we can hope to "mod out" the issue in some way)

HHansKristian-Work 2025-01-24 github

game workable on GTX 16XX series GPUs

The game hard requires mesh shaders, and will error if FL 12.2 is not supported. I guess a mod that pretends 12.2 is supported on a non-RTX Turing will in theory work.

The bug in question related to mesh shaders being bugged on NV.

Do you believe such a "mod" ...

No. This needs driver fixes.

HHansKristian-Work 2025-01-24 github

I use proton experimental bleeding edge and i still see white flashbangs full screen on mesa RADV drivers doesnt seem to happen on AMD pro drivers. (RX 6700 XT)

Do you have a proton log proving that? It's possible it ends up not using experimental somehow.

CCha1N1 2025-01-24 github

I use proton experimental bleeding edge and i still see white flashbangs full screen on mesa RADV drivers doesnt seem to happen on AMD pro drivers. (RX 6700 XT)

Do you have a proton log proving that? It's possible it ends up not using experimental somehow.

it doesn't let me upload anything so im sending what the proton log spammed over and over:
8507.193:0138:02f0:warn:vkd3d-proton:d3d12_pipeline_library_load_pipeline: Pipeline "EAA5EF7558AC200F" does not exist.
8507.194:0138:02f8:warn:vkd3d-proton:d3d12_pipeline_library_load_pipeline: Pipeline "9B9CC9D2CBBF38F3" does not exist.
8507.195:0138:02f4:warn:vkd3d-proton:d3d12_pipeline_library_load_pipeline: Pipeline "854ABC93A01E9DD8" does not exist.
8507.196:0138:0310:warn:vkd3d-proton:d3d12_pipeline_library_load_pipeline: Pipeline "32CFE69502D8960A" does not exist.
8507.196:0138:0300:warn:vkd3d-proton:d3d12_pipeline_library_load_pipeline: Pipeline "C608EA2E06D49C31" does not exist.
8507.196:0138:0304:warn:vkd3d-proton:d3d12_pipeline_library_load_pipeline: Pipeline "C2B94A59E98B38BF" does not exist.
10827.688:0138:0774:warn:vkd3d-proton:d3d12_command_queue_QueryInterface: {db6f6ddb-ac77-4e88-8253-819df9bbf140} not implemented, returning E_NOINTERFACE.
10827.693:0138:0224:warn:vkd3d-proton:vkd3d_native_sync_handle_release: Failed to release semaphore (#12a).
10827.693:0138:0224:warn:vkd3d-proton:dxgi_vk_swap_chain_wait_worker: Failed to increment swapchain semaphore. Did application forget to acquire?

Sadly i cant attach images cause my internet might be acting up right now upload fails.

Edit: could upload an image of it on an external site if its okay https://imgur.com/a/1ofiaKY

Aandrew-ld 2025-01-24 github

gpu 7900xtx
cpu 7800x3d
os archlinux
4k@144hz
mesa-git
proton-ge 9-23

it seems that when you move camera view there is an abnormal drop in fps even if the gpu is not having a high workload, to almost completely solve the problem i did the following things:

i started the game with
VKD3D_CONFIG=small_vram_rebar,no_staggered_submit gamemoderun gamescope --adaptive-sync -h 2160 -r 144 -f -- %command%

reason:

  • it seems that the binary name does not end with -shipping.exe, this prevents vkd3d from applying the workarounds that are there by default for all unreal engine games, so I applied them with vkd3d_config.
  • It seems in addition that starting the game with gamescope and gamemoderun seems to improve a bit more.

I changed the following settings in the game:
graphics quality: high
minimum/maximum upscale: 66%
maximum framerate: 90fps

i put maximum 90fps because it seems that my gpu most of the time can hold this framerate maximum and going beyond it starts to vary the number of fps too often.

in addition I used https://www.nexusmods.com/site/mods/757 to be able to use intel xess as an upscaler

now the game is quite smooth even moving the camera view

Sslbillups 2025-01-24 github

i started the game with VKD3D_CONFIG=small_vram_rebar,no_staggered_submit gamemoderun gamescope --adaptive-sync -h 2160 -r 144 -f -- %command%

The 'shipping' executable you're used to seeing is located in the End/Binaries/ directory. I'm not going to try to look into the root directory executable for fear of the SE gods bringing the hammer down. However, even if you attempt to launch from the baked executable you won't be able to launch it from the path directly - if you do you'll get the FF7 splash screen with a message saying to start the game from the Steam client

Image

to almost completely solve the problem i did the following things:

By problems are you referring to the textures failing to load in when the scene initially loads e.g. the title of this issue? Or is this just a general performance tweak for better FPS on your AMD card?

This issue I thought was exclusive to Nvidia GPUs.

inxi -b
System:
Host: Kernel: 6.13.0-2-cachyos arch: x86_64 bits: 64
CPU:
Info: 6-core Intel Core i7-8700K [MT MCP] speed (MHz): avg: 4401
min/max: 800/4700
Graphics:
Device-1: NVIDIA GA106 [GeForce RTX 3060 Lite Hash Rate] driver: nvidia
v: 565.77
Display: wayland server: X.org v: 1.21.1.15 with: Xwayland v: 24.1.4
compositor: kwin_wayland driver: X: loaded: nvidia unloaded: modesetting
gpu: nvidia,nvidia-nvswitch resolution: 1920x1080~240Hz
API: OpenGL v: 4.6.0 vendor: nvidia v: 565.77 renderer: NVIDIA GeForce
RTX 3060/PCIe/SSE2
Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo
de: kscreen-console,kscreen-doctor gpu: nvidia-settings,nvidia-smi
wl: wayland-info x11: xdpyinfo, xprop, xrandr
Info:
Memory: total: 32 GiB available: 31.28 GiB used: 4.04 GiB (12.9%)

Oofcaah 2025-01-24 github

This issue I thought was exclusive to Nvidia GPUs.

it is, but some AMD guys decided to squat here and didn't get the subtle messaging to go away, and build their own issue ;)
that being said, original issue seems to be resolved with 'wait for new drivers and hope for the best'.

Zzr00da 2025-01-24 github

it is, but some AMD guys decided to squat here and didn't get the subtle messaging to go away, and build their own issue ;) that being said, original issue seems to be resolved with 'wait for new drivers and hope for the best'.

Issues are filed per game, not per some specific one issue. Keep this edgy gatekeeping for somewhere else please.

Oofcaah 2025-01-24 github

Issues are filed per game, not per some specific one issue

If so, then issue template requiring specifying the GPU, describing the symptoms, logs etc. seems misleading, not to mention the added "NVIDIA Drivers" label.
And nobody's gatekeeping anything. I merely suggested, basing on the above, that creating a second issue regarding the AMD problems might just be "cleaner" for everyone. If that's only me, so be it - I won't be mentioning this again.

Aandrew-ld 2025-01-24 github
[andrew@arch FINAL FANTASY VII REBIRTH]$ pwd
/home/andrew/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH
[andrew@arch FINAL FANTASY VII REBIRTH]$ find . -name "*.exe"
./End/Binaries/Win64/ff7rebirth_.exe
./End/Binaries/Win64/unins000.exe
./End/Binaries/Win64/dlss-finder.exe
./ff7rebirth.exe

@slbillups I don't have any binary that has -shipping


cyberpunk issue 1700 comments, gta v issue 1600 comments, i dont see how for a couple of comments we have to isolate this issue to nvidia problems

Sslbillups 2025-01-24 github

Because this is a nvidia specific issue that is related to the dxgi swapchain/VkPipeline failing to stream in and render the world textures. I don't frequent this repository, and since the title of the post wasn't titled with anything mentioning that this is a nvidia-specific issue, the nvidia developer forums or the VKD3D repo may be a better place for discussion.

However when it comes to games, I see many users who either create a new github account for the sole purpose of adding information that is likely just an issue with the end-user hardware configuration. I think these issues are more suited for the Steam discussion platform, protondb, etc rather than the Proton github repository...

Rroccodev 2025-01-24 github

@andrew-ld I tried your vkd3d flags and they didn't seem to make that much of a difference, the first one in particular is supposed to only work with cards with <= 8GB VRAM.

Though as I mentioned in my first post, I think what really helps there is disabling the dynamic resolution scaler, which you did by setting the range to 66-66%. I also think that the LOD pop-in issue I was experiencing is due to me running the game at 4K (I had a range of 66-100% res scale, and setting it to 100-100% made it more consistent), I might be hitting some VRAM limits there, so it makes sense. At 66-66% the pop-in issue goes away.

SSloogs 2025-01-25 github

Confirming I am also experiencing the missing textures issue on a 3060 Ti, Fedora 41, driver 565.77, kernel 6.12.10-200.fc41.x86_64

Jjeffhirsch 2025-01-26 github

i started the game with
VKD3D_CONFIG=small_vram_rebar,no_staggered_submit gamemoderun gamescope --adaptive-sync -h 2160 -r 144 -f -- %command%

not sure if it was the game update that was released today or these flags but it is working great for me now. Thanks @andrew-ld

5800x3d
7900xtx
6.12.10-203.nobara.fc40.x86_64

Aandrew-ld 2025-01-26 github

@jeffhirsch I tried again today and it looks like the VKD3D_CONFIG workaround is still needed

Jjolty1 2025-01-26 github

I get this error using my DS4 Official Sony Wireless Adaptor. Tried Disabling/Enabling Steam Input.

Works fine using Wired.

Bus 001 Device 008: ID 054c:0ba0 Sony Corp. Dualshock4 Wireless Adaptor

LowLevelFatalError [File:Unknown] [Line: 103] 
SECURE CRT: Invalid parameter detected.
Expression: Unknown Function: Unknown. File: Unknown Line: 0



0x00006fffffc1cf57 kernelbase.dll!UnknownFunction []
0x000000014463533f ff7rebirth_.exe!UnknownFunction []
0x0000000144636c94 ff7rebirth_.exe!UnknownFunction []
0x000000014246ff03 ff7rebirth_.exe!UnknownFunction []
0x00006fffffa6508b ucrtbase.dll!UnknownFunction []
0x0000000141b6cee6 ff7rebirth_.exe!UnknownFunction []
0x00006fffffa0111c ucrtbase.dll!UnknownFunction []
0x00006fffffed482d kernel32.dll!UnknownFunction []

Proton Log if it helps.

steam-2909400.log

Zzixaphir 2025-01-26 github

This game is steam deck verified. Why is this labeled as "unofficial"?

Lliberodark 2025-01-26 github

Hi,

Other issue is on AMD 7900 XTX with RADV (24.3.4) & Kernel 6.12.10 :

Image

Image

Qq4niel 2025-01-26 github

So I guess I have no choice but to boot up Windows for now, hope this gets resolved soon.

Issue:
Various textures not rendering. Some Nvidia driver problem as far as my understanding goes.

OS: Arch Linux
Compositor: Hyprland
GPU: RTX 3070 Ti
GPU DRIVER: nvidia-dkms
CPU: Ryzen 5 5600X

Zzr00da 2025-01-26 github

@liberodark As mentioned here https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2612790354 that's fixed by using Proton Experimental

SSheMelody 2025-01-27 github

Happens to my boyfriend using a 4080 SUPER, both NVIDIA Open 565.77 and NVIDIA Open 570.86, Arch Linux. Using Proton Experimental Bleeding Edge. Tried both SteamDeck=0 and SteamDeck=1.

Let us know what we can do to help fixing this problem.

Oofcaah 2025-01-27 github

Let us know what we can do to help fixing this problem.

See here:
https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2613018177

TL;DR: Wait for next nvidia drivers and hope for the best.

Oofcaah 2025-01-27 github

in case anyone's wondering: no, 570.86.15 that went public today do not resolve this issue. :(

edit:
TIL: frown/meh emoji hides comment as off-topic ;)

WWhitleyStriber 2025-01-28 github

Also having the same issue.

Latest Arch Linux
EVGA 3080 Ti

Mmmxgn 2025-01-28 · hidden on GitHub github

in case anyone's wondering: no, 570.86.15 that went public today do not resolve this issue. :(

From what one can read these are Datacenter GPUs, I would not expect them to fix gaming issues.
https://docs.nvidia.com/datacenter/tesla/pdf/NVIDIA_Data_Center_GPU_Driver_Release_Notes_570_v1.0.pdf

SSheMelody 2025-01-28 · hidden on GitHub github

in case anyone's wondering: no, 570.86.15 that went public today do not resolve this issue. :(

From what one can read these are Datacenter GPUs, I would not expect them to fix gaming issues. https://docs.nvidia.com/datacenter/tesla/pdf/NVIDIA_Data_Center_GPU_Driver_Release_Notes_570_v1.0.pdf

This is not entirely true. This new driver actually boosts VKD3D performance as well as native Vulkan performance by a good ton. The difference is definitely there and you can clearly see it.

Now, the issue with NVIDIA and drivers is that they mainly focus on AI (tensorrt, pytorch, onnx and so forth), yet this does not mean they don't care at all about Vulkan and OpenGL, it just means they don't care enough.

It is an undeniable fact that most NVIDIA Linux machines are usually used for AI workflows, ie. Google colab, remote GPU datacenter solutions in general, or just private machines running AI. If you compare AI performance on Linux vs Windows on NVIDIA, Linux completely destroys Windows, even on old drivers and LTS kernels.

Maybe a company like Valve can make gaming look a bit more important to NVIDIA. Users alone probably can't really do much. My guess is that gamers running Linux are an extremely tiny minority to NVIDIA and they don't see a real incentive compared to improving AI.

PS tensorrt does not even work properly on Windows for me.

SSheMelody 2025-01-28 github

Let us know what we can do to help fixing this problem.

See here: [#8408 (comment)](https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2613018177)

TL;DR: Wait for next nvidia drivers and hope for the best.

In Windows, we have tried hardcapping feature level for the game to both 12_1 and 12_0 while also disabling Feature Level Upgrade altogether and the game still looks fine despite running at lower framerate. I do not think this is driver-related, but feel free to change my mind.

Zzixaphir 2025-01-28 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2618892110

The quoted post below is by Hans-Kristian Arntzen, the creator of vkd3d-proton. I assume the workaround he was trying to create would be within vkd3d.

The RTX issues are known. It's a driver bug and it has been reported to NV with bug number 5059946.

I've spent a lot of time trying to figure out a reasonable workaround, but to no avail.

SSheMelody 2025-01-28 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2619048268

What does this bug do exactly? I thought the problem was mesh shaders, right?

Zzixaphir 2025-01-28 github

What does this bug do exactly? I thought the problem was mesh shaders, right?

I'm not a technical enough person to understand the actual issue at play

DDistantThunder 2025-01-28 github

Doc - Fix Resolution issue within Gamescope

gamescope --force-windows-fullscreen -b -w $(kscreen-doctor -j | jq '.screen.currentSize.width') -h $(kscreen-doctor -j | jq '.screen.currentSize.height') -H $(kscreen-doctor -j | jq '.screen.currentSize.height') --hdr-enabled --mangoapp -- %command%

The --force-windows-fullscreen is key.

Eelken 2025-01-28 github

Those having the issues with full system crashes on AMD cards like me should look at https://gitlab.freedesktop.org/drm/amd/-/issues/3131

I used to have to play on low settings in a tiny window to prevent the high usage but now I can play maxed out for as long as 10 minutes.

Will have to play test longer but this might help some. Short version is the clock speeds for the GPU are higher than what should be set by the manufacturer so either set them manually or try setting the profile daemon to manual then back to auto

Aandrew-ld 2025-01-29 github

@elken i am the author of the issue on free desktop, i think that in this case the GPU crash is because of a bug in the game, vkd3d Is trying to workaround the issue

maybe is even crashing for that issue but anyway there is another crash caused by the game

SSheMelody 2025-01-29 github

Since there's people thumbing me down all the time in this issue, I'd like to elaborate on my troubleshooting process.

  • I was told the problem with this issue was linked to mesh shaders
  • -> so I tried making them unavailable in Windows, which still didn't cause the issue to happen, or any issue whatsoever, hence why I'm personally not convinced it's a driver issue

If straight reducing the available shader model version is not enough to make this issue appear, so how does this bug work exactly?

Edit: On a side note, reducing FL to 12_1 or even to 12_0 also causes even more features to be unavailable. It's weird that the game is still able to work properly with Direct3D 12 FL 12_0 assuming that missing mesh shaders would completely break terrain and most of the scenery.

I'd prefer receiving a constructive answer on this matter rather than thumbs down if that's possible, thank you.

Eelken 2025-01-29 github

@elken i am the author of the issue on free desktop, i think that in this case the GPU crash is because of a bug in the game, vkd3d Is trying to workaround the issue

maybe is even crashing for that issue but anyway there is another crash caused by the game

The issue in question fixed all my issues, and I've seen others having the exact same kind of crash as me 😄

Nneliavaal 2025-01-29 github

vkd3d-proton-pso-retain-references.zip which contains this workaround and hopefully helps with the AMD driver crash (I suppose).

Oofcaah 2025-01-29 github
* I was told the problem with this issue was linked to **mesh shaders**

* -> so I tried **making them unavailable in Windows**, which **still didn't cause the issue to happen**, or any issue whatsoever, hence why I'm personally not convinced it's a driver issue

My guess would be:

  1. the problem occurs when mesh shaders are available, but broken. (as is on currently available NV drivers up to and including 570.86.15)
  2. game checks for available features once during startup, pops an error message and quits if hard requirements are not met, and doesn't bother with this further down the line, therefore blindly using "unavailable" features anyway, as long as initial checks were satisfied.
SSheMelody 2025-01-29 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2621728724

A quick workaround would be disabling mesh shaders altogether, since feature level 12_0 theoretically don't expose mesh shaders even if available, right? FF seems to run when forcing it to FL 12_0 on Windows, without mesh shaders, it looks fine but it runs at a lower framerate.

HHansKristian-Work 2025-01-29 github

FF seems to run when forcing it to FL 12_0 on Windows, without mesh shaders, it looks fine but it runs at a lower framerate.

Could you explain how you did that? Because the game refuses to start if it cannot create a FL 12.2 device when I tested.

SSheMelody 2025-01-29 github

FF seems to run when forcing it to FL 12_0 on Windows, without mesh shaders, it looks fine but it runs at a lower framerate.

Could you explain how you did that? Because the game refuses to start if it cannot create a FL 12.2 device when I tested.

  • Download and install Graphics Tools from Optional Features
  • Open dxcpl from a Run box
  • Add FF's executable
  • Cap feature level and make sure to disable Feature Level Upgrade
  • Apply
Oofcaah 2025-01-29 github
* Cap feature level and make sure to disable Feature Level Upgrade

Assuming this works properly, and attempts to use mesh shaders are indeed blocked , then perhaps it's the other way around: game checks for FL 12.2, is being lied to that it's supported, and then Unreal Engine manages to work around missing features? That would explain how people made FF work on GTX 1650, even though it obviously doesn't support 12.2

Can the same be attempted with vkd3d? i.e. reporting 12.2 available, but then failing attempts to use 12.2 features, or at least mesh shaders?

HHansKristian-Work 2025-01-29 github

It's not impossible. I haven't tried that myself at least.

SSheMelody 2025-01-29 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2621817808

Assuming the game attempts to create FL 12_2 and cannot fallback, then yes, it is being lied to and then Unreal Engine works missing features around. That's plausible.

HHansKristian-Work 2025-01-29 github

I tried this, but the game fails to boot with an error about missing assets:

diff --git a/libs/vkd3d/device.c b/libs/vkd3d/device.c
index 494bb067..201f0af6 100644
--- a/libs/vkd3d/device.c
+++ b/libs/vkd3d/device.c
@@ -4303,6 +4303,8 @@ static HRESULT STDMETHODCALLTYPE d3d12_device_CheckFeatureSupport(d3d12_device_i
                     data->MaxSupportedFeatureLevel = fl;
             }
 
+            data->MaxSupportedFeatureLevel = D3D_FEATURE_LEVEL_12_0;
+
             TRACE("Max supported feature level %#x.\n", data->MaxSupportedFeatureLevel);
             return S_OK;
         }
@@ -4702,6 +4704,7 @@ static HRESULT STDMETHODCALLTYPE d3d12_device_CheckFeatureSupport(d3d12_device_i
             }
 
             *data = device->d3d12_caps.options7;
+            data->MeshShaderTier = 0;
 
             TRACE("Mesh shading tier %#x.\n", data->MeshShaderTier);
             TRACE("Sampler feedback tier %#x.\n", data->SamplerFeedbackTier);
SSheMelody 2025-01-29 github

I have no idea how dxcpl manages to boot the game up with tier 0 mesh shaders (FL 12_0). Is it plausible to assume that Windows would be emulating them through WARP (not to be confused with "force warp" which forces software rendering for everything)? Maybe it's not UE compensating, it could be Direct3D itself.

Edit: This would also explain why the game runs 15-20 fps slower when forcing 12_0.

Zzixaphir 2025-01-29 · hidden on GitHub github

I have no idea how dxcpl manages to boot the game up with tier 0 mesh shaders (FL 12_0). Is it plausible to assume that Windows would be emulating them through WARP (not to be confused with "force warp" which forces software rendering for everything)? Maybe it's not UE compensating, it could be Direct3D itself.

Edit: This would also explain why the game runs 15-20 fps slower when forcing 12_0.

Hey thank you for ignoring me earlier. I'm over here running my mouth and you're getting actual work done. Gotta learn to shut my face lol

HHansKristian-Work 2025-01-29 github

Edit: This would also explain why the game runs 15-20 fps slower when forcing 12_0.

I think we need some way to incontrovertibly prove that the game is actually demoting mesh shaders to plain-old vertex shaders here somehow. The lowered FPS is curious, but it's unclear what the dxcpl tool ends up doing.

Mmatezoltanfarkas 2025-01-29 github

Edit: This would also explain why the game runs 15-20 fps slower when forcing 12_0.

Can you reproduce these results with other games like Alan Wake 2 as well?
I still do not understand why mesh shaders are the issue here with nvidia users. Alan Wake 2 used them extensively, and I had no issues with it -- the game did not even warn me about missing features whatsoever.

SSheMelody 2025-01-29 github

Edit: This would also explain why the game runs 15-20 fps slower when forcing 12_0.

I think we need some way to incontrovertibly prove that the game is actually demoting mesh shaders to plain-old vertex shaders here somehow. The lowered FPS is curious, but it's unclear what the dxcpl tool ends up doing.

Yeah, I agree, but I don't know where to start to begin with. Unlucky.

Edit: This would also explain why the game runs 15-20 fps slower when forcing 12_0.

Can you reproduce these results with other games like Alan Wake 2 as well? I still do not understand why mesh shaders are the issue here with nvidia users. Alan Wake 2 used them extensively, and I had no issues with it -- the game did not even warn me about missing features whatsoever.

I unfortunately don't own Alan Wake 2. But yes, you're right, Alan Wake 2 actually has mesh shaders as a strict requirement.

Mmatezoltanfarkas 2025-01-29 github

maybe benchmarking the nvidia asteroids demo might work as well? link

I have tried to get it running in bottles, but it closes on my machine with a "mesh shaders not supported, demo will close"-like error message -- despite the lines below:

11068.130:0020:0024:info:vkd3d-proton:d3d12_device_determine_ray_tracing_tier: DXR support enabled.
11068.130:0020:0024:info:vkd3d-proton:d3d12_device_determine_ray_tracing_tier: DXR 1.1 support enabled.
11068.130:0020:0024:info:vkd3d-proton:d3d12_device_caps_init_feature_level: DX Ultimate supported!

SSheMelody 2025-01-29 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2622223633

If you can, try capping FL to 12_0 under Windows for the executable and see if it still runs without complaining. I can't test right now.

HHansKristian-Work 2025-01-29 github

No need to try that. I verified that the dxcpl tool thingy does not do anything for d3d12, and also confirmed that with someone who knows the details. That override is for d3d11 only. I also verified through capturing the game on Windows that it is still using mesh shaders.

SSheMelody 2025-01-29 github

No need to try that. I verified that the dxcpl tool thingy does not do anything for d3d12, and also confirmed that with someone who knows the details. That override is for d3d11 only. I also verified through capturing the game on Windows that it is still using mesh shaders.

Uhm, this is what I get by capping Feature level limit + disable feature level upgrade on Cyberpunk 2077:

Image

It says there has been an error initializing Ray Tracing and that my drivers could be obsolete or damaged. It's telling me to reinstall my drivers.

Are you sure you're not using the dxcpl from the old DirectX SDK? Make sure it says Direct3D 10/11/12:

Image

You can get the updated dxcpl from PC settings -> Optional Features

(regardless, a quick google search confirms that dxcpl works for d3d12, other than this I have no idea how to help)

Edit: sometimes dxcpl doesn't work, see my new answer below

LLethalManBoob 2025-01-29 · hidden on GitHub github

No need to try that. I verified that the dxcpl tool thingy does not do anything for d3d12, and also confirmed that with someone who knows the details. That override is for d3d11 only. I also verified through capturing the game on Windows that it is still using mesh shaders.

buds trying to dodge work

RRevengeRip 2025-01-29 github

I filed a bugreport at mesa for amdgpu crashes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12545

SSheMelody 2025-01-29 github

Small correction: It seems like you were right about the game still using mesh shaders, dxcpl doesn't prevent the game from using them. Welp, no idea what's going on honestly then, Direct3D works in weird ways.

People still use dxcpl in combination with D3D12 to change the feature level. It seems like some games can forcibly decide what to do, somehow. This is not always the case for all games. CP2077 really really didn't like me changing the FL limit.

Edit: Ratchet & Clank: Rift Apart also reacts in a bad way, but only after going through the launcher and launching the game.

Oofcaah 2025-01-29 github

I've managed to replicate Hans' error; also tried numerous more forced "downgrades", and each attempt resulted either in DX12 being unsupported error, "Game files required to initialize the global shadre library are missing" error, or just generic "Fatal error!" while "checking saved data..."

That aside, has anyone tried 550.40.83 drivers?

Vulkan Developer Beta Driver Release Updates
January 24th, 2025 - Windows 553.64, Linux 550.40.83

    New:
        [VK_KHR_depth_clamp_zero_one](https://registry.khronos.org/vulkan/specs/latest/man/html/VK_KHR_depth_clamp_zero_one.html)
        [VK_KHR_video_maintenance2](https://registry.khronos.org/vulkan/specs/latest/man/html/VK_KHR_video_maintenance2.html)
    Fixes:
        Allow linear images with VK_IMAGE_CREATE_DESCRIPTOR_BUFFER_CAPTURE_REPLAY_BIT_EXT
        > **Fix gl_NumWorkGroups behavior for mesh shaders** <
        Fix alignment requirement for buffers created with VK_BUFFER_USAGE_2_PREPROCESS_BUFFER_BIT_EXT
        Don't attempt to clear fragment shading rate attachments used with dynamic rendering

EDIT: Just tested them; unfortunately nothing has changed. Still no ground post intro sequence.

Jjorbuedo 2025-01-29 github

At least they're looking into it internally at nvidia https://github.com/NVIDIA/open-gpu-kernel-modules/issues/708#issuecomment-2623167940

Pity their bug reports can't really be viewed so no visibility on the progress.

HHansKristian-Work 2025-01-30 github

For AMD hangs, we have two game workarounds merged now which have successfully worked around some hangs on our end at least:

Oofcaah 2025-01-30 github

Nvidia released 570.86.16 today: https://www.nvidia.com/en-us/geforce/drivers/results/240524/

Sadly, it doesn't resolve the problem. Oh well.

Zzixaphir 2025-01-30 · hidden on GitHub github

Nvidia released 570.86.16 today: https://www.nvidia.com/en-us/geforce/drivers/results/240524/

Sadly, it doesn't resolve the problem. Oh well.

Unfortunately despite Nvidia-open Nvidia's internal development process is still super opaque. Unless one of their employees decides to be open about where development is on a fix, we'll likely be in the dark except for the internal bug only the filer can see. Here's hoping things will get more transparent at least for the driver eventually.

RRevengeRip 2025-01-30 github

For AMD hangs, we have two game workarounds merged now which have successfully worked around some hangs on our end at least:

* [Add PSO retain workaround for FFVII Rebirth. HansKristian-Work/vkd3d-proton#2319](https://github.com/HansKristian-Work/vkd3d-proton/pull/2319)

* [Workaround missing CS barrier in Rebirth. HansKristian-Work/vkd3d-proton#2320](https://github.com/HansKristian-Work/vkd3d-proton/pull/2320) (from 1 hour ago)

Thanks, those really solve gpu hangs for me!

Mmmequignon 2025-01-30 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2625433490

I'm sorry but, how did you test it? Do I have to compile vkd3d-proton or something like that? Or maybe using proton experimental is enough to have the latest version?
I don't know much about this…

Nneliavaal 2025-01-30 github

@HansKristian-Work works for me! THANK YOU!

@mmequignon

vkd3d-proton-pso-retain-references.zip

which contains the mentioned workarounds. Replace the dlls inside proton and make sure to use a clean prefix.
mangohud will show 99.0.0 vkd3d version

Mmmequignon 2025-01-30 github

@HansKristian-Work works for me! THANK YOU!

@mmequignon

vkd3d-proton-pso-retain-references.zip

which contains the mentioned workarounds. Replace the dlls inside proton and make sure to use a clean prefix. mangohud will show 99.0.0 vkd3d version

I'm sorry, I don't quite understand what you mean. Could you please give me a hand? I'm quite the beginner here.
If you're ok, you can reach me by email, it is on my gh account.
Thanks!

HHansKristian-Work 2025-01-30 github

Just use Experimental bleeding-edge. There is no need to download dlls like this. We're in the process of pinning a Hotfix build that should automate this stuff.

Mmmequignon 2025-01-31 github

Just use Experimental bleeding-edge. There is no need to download dlls like this. We're in the process of pinning a Hotfix build that should automate this stuff.

Ah great, I didn't know about that, thank you. 🙇
I'll try this tomorrow and give feedback!

Ccyberiacafe 2025-01-31 github

Any news for Nvidia users on the issue we're having? Is it even a Nvidia issue considering a new driver came out yesterday and the issue persisted.

UUklosk 2025-01-31 github

Any news for Nvidia users on the issue we're having? Is it even a Nvidia issue considering a new driver came out yesterday and the issue persisted.

NVIDIA just don't care about us Linux users, NVK is our only hope in the long term.

OOkerlund 2025-01-31 github

Any news for Nvidia users on the issue we're having? Is it even a Nvidia issue considering a new driver came out yesterday and the issue persisted.

On the Nvidia website it is listed as a beta driver. They have mentioned they are looking into the issue, so hopefully by time, or before, it comes out of beta it will be fixed.

Oofcaah 2025-01-31 · hidden on GitHub github

NVIDIA just don't care about us Linux users, NVK is our only hope in the long term.

I've been playing games on Linux on an RTX 3090 for three years now, and their proprietary drivers have pretty much worked. They also made Wayland support work recently, so saying they don't care about Linux users is simply unfair, and quite an overstatement.

NVK still lacks features needed to support FL 12_2 (even on paper); one of those missing features is... mesh shaders. Work is being done, and the day FF7 Rebirth runs under NVK will surely be a considerable milestone and a testament to NVK's viability.

Less than 2.5% players on Steam are currently Linux users. Roughly a third of that number are Steam Deck owners (AMD) and in total, about 15% of that ~2.5% of Linux users have an NVIDIA RTX GPU. Expecting NVIDIA to immediately devote all its resources to a niche bug in a relatively niche game on a niche platform (mostly dominated by AMD GPUs), especially amid the hot timeframe of RTX 50XX release, is simply too tall an order.

If you really can't wait, do what I did: grab a spare 1TB NVMe, install some Windows LTSC and you'll be playing within an hour or two. Your saves will transfer. Enjoy the weekend! :)

UUklosk 2025-01-31 · hidden on GitHub github

NVIDIA just don't care about us Linux users, NVK is our only hope in the long term.

I've been playing games on Linux on an RTX 3090 for three years now, and their proprietary drivers have pretty much worked. They also made Wayland support work recently, so saying they don't care about Linux users is simply unfair, and quite an overstatement.

NVK still lacks features needed to support FL 12_2 (even on paper); one of those missing features is... mesh shaders. Work is being done, and the day FF7 Rebirth runs under NVK will surely be a considerable milestone and a testament to NVK's viability.

Less than 2.5% players on Steam are currently Linux users. Roughly a third of that number are Steam Deck owners (AMD) and in total, about 15% of that ~2.5% of Linux users have an NVIDIA RTX GPU. Expecting NVIDIA to immediately devote all its resources to a niche bug in a relatively niche game on a niche platform (mostly dominated by AMD GPUs), especially amid the hot timeframe of RTX 50XX release, is simply too tall an order.

If you really can't wait, do what I did: grab a spare 1TB NVMe, install some Windows LTSC and you'll be playing within an hour or two. Your saves will transfer. Enjoy the weekend! :)

Expecting a working driver is now too much to ask, apparently.

Oofcaah 2025-01-31 · hidden on GitHub github

Expecting a working driver is now too much to ask, apparently.

Perhaps manage your expectations? And if you are really into shifting blame around, then why was Hans the one filing the bug report with Nvidia, and not Square Enix? It's not some obscure bug that becomes apparent after 20 hours of intensive play testing. Literally launching the game on Linux with an RTX would expose this problem months before release, so it simply shows that they did not do that even once, or they did and moved on. Who doesn't care about "us Linux users" now?

Zzixaphir 2025-01-31 github

Please stop arguing about Nvidia. I'm following this issue for an update on when I can play this game, not to have my inbox spammed with arguments about how good or bad they are as a company. Please, if you wanna argue about Nvidia on Linux, go to Reddit or something.

Aarnxxau 2025-01-31 github

I'm experiencing massive stutters just by moving the camera around.
I tried doing the same on Windows, and it doesn't happen.

I also tried this launch command and disabled dynamic res

Additionally, I set the graphics to low and capped the framerate at 30 FPS, but the issue still persists.


gpu 6700xt
cpu 5700x3d
os archlinux
linux zen kernel
mesa-git
proton-ge 9-23 and proton experimental bleeding edge

Ccyberiacafe 2025-01-31 github

Any news for Nvidia users on the issue we're having? Is it even a Nvidia issue considering a new driver came out yesterday and the issue persisted.

On the Nvidia website it is listed as a beta driver. They have mentioned they are looking into the issue, so hopefully by time, or before, it comes out of beta it will be fixed.

Sweet, thanks for the info, suppose i'll replay the first remake in the meantime

Ee88z4 2025-02-01 github

For AMD hangs, we have two game workarounds merged now which have successfully worked around some hangs on our end at least:

I can confirm these patches fixed the lighting issue. I use proton GE 9-23 and override the vkd3d dll with the latest master commit on de2987b

The lighting was too bright but it is fixed with the patch.
Image

The vending machine was flickering and transparent previously. After the patch, the object is rendered correctly.

Image

I am running Debian unstable, Ryzen 5 5600x, AMD Radeon 7800XT. No crash anymore so far with my limited testing.

Jjeffhirsch 2025-02-01 github

Additionally, I set the graphics to low and capped the framerate at 30 FPS, but the issue still persists.

gpu 6700xt
cpu 5700x3d

@arnxxau make sure you're on proton experimental bleeding edge as referenced here. That fixed all my issues on amd.

Eejrydhfs 2025-02-01 github

The RTX issues are known. It's a driver bug and it has been reported to NV with bug number 5059946.
I've spent a lot of time trying to figure out a reasonable workaround, but to no avail.

Is there a link to the bug report so we can track the progress? Also I'm experiencing the exact same issue, RTX 3060Ti + 565.77.

I am afraid no. I have reported bugs to NV in the past and their bugs are kept private to NV and the bug reporter and this is enforced by only using email for comms

Aarnxxau 2025-02-01 github

Additionally, I set the graphics to low and capped the framerate at 30 FPS, but the issue still persists.

gpu 6700xt
cpu 5700x3d

@arnxxau make sure you're on proton experimental bleeding edge as referenced here. That fixed all my issues on amd.

Hi @jeffhirsch, I'm sure I was on Proton Experimental Bleeding Edge. The lighting issues and transparent objects were fixed, but I'm still experiencing a stuttering behavior when moving the camera around, which does not happen on Windows.

I tested it on Windows to see how bad the stutters really were since reviews mentioned that stuttering does occur. However, when I tried it on Windows, I realized that the stuttering wasn't as bad as I expected, and the way it currently runs on Linux doesn't seem to be the correct behavior.

I also tried:

  • Limiting the FPS with external limiters
  • Changing to various proton versions (proton9, protonge, proton experimental.....)
  • Changing my kernel to amd-drm-next and CachyOS
  • Running the game through Gamescope
  • Running the entire Steam session through Gamescope (like on the Steam Deck)
  • Disabling VRR in-game, within GNOME, and in Gamescope
  • Disabling VKD3D cache
  • Disabling DirectStorage with FF7Hook
  • Updating to the latest mesa git

None of these fixed the stuttering.

Mmmequignon 2025-02-01 github

Just use Experimental bleeding-edge. There is no need to download dlls like this. We're in the process of pinning a Hotfix build that should automate this stuff.

Just tried to run the game with experimental bleeding edge, and I'm still experiencing random crashes.

CPU: AMD Ryzen 7 3700X
GPU: AMD RX 7900 XT
Kernel is 6.12.10

OOkerlund 2025-02-01 github

Square Enix has made a post on the Steam forums that they are investigating issues with the game when using AMD. https://steamcommunity.com/app/2909400/discussions/0/597391264850139769/

CCha1N1 2025-02-01 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2628946388

i exhibit the same stuttering on my system
With an RX 6700 XT and 5600X on 6.13 Cachyos kernel
it only happens when i turn the camera otherwise experimental bleeding edge fixed all pain points for me, but the stutter is really jarring and can't be fixed even with fps cap. One thing to note is when i updated the game's direct storage version, it seemed to improve quite a bit, but when the camera moves my cpu gets extremely hammered. Possibly IO read speed issues or something can't keep up?

EEoinHealy0 2025-02-03 github

Todays Rebirth patch appears to do nothing to help with the Nvidia issue.

OOkerlund 2025-02-03 github

Just got the Nvidia 570 drivers on Arch today and ground textures still don't load. And can confirm that the patch today also did nothing to fix it.

OS: Arch Linux
KERNEL: 6.13.1-arch1-1
CPU: 13th Gen Intel Core i9-13900K
GPU: NVIDIA GeForce RTX 4080
GPU DRIVER: NVIDIA 570.86.16

Ppurplekairi1312 2025-02-03 · hidden on GitHub github

Updated to last nvidia driver and can confirm too that game ground textures still won't load

OS: Arch Linux
KERNEL: 6.13.1-arch1-1 and 6.12.12-1-lts
CPU: i9-14900K
GPU: NVIDIA GeForce RTX 3090
GPU DRIVER: 570.86.16

Sspiffeeroo 2025-02-04 github

I'm experiencing massive stutters just by moving the camera around. I tried doing the same on Windows, and it doesn't happen.

I also tried this launch command and disabled dynamic res

Additionally, I set the graphics to low and capped the framerate at 30 FPS, but the issue still persists.

gpu 6700xt cpu 5700x3d os archlinux linux zen kernel mesa-git proton-ge 9-23 and proton experimental bleeding edge

Try updating to Linux kernel 6.13.1 if you have an AMD GPU. In Linux kernel 6.12.10 and older, the power saving was too aggressive on my AMD graphics cards (RX 7900 XTX and RX 7800 XT). The VRAM P-state clock speed were stuck between 96 MHz, 456 MHz, and 772 MHz instead of the max of 1249 MHz. Linux kernel 6.13.1 on Arch Linux has the VRAM clock speed go up to 1249 MHz as expected in this game.

You can see the VRAM clock speed in MangoHud or if you run cat /sys/class/drm/card1/device/pp_dpm_mclk.

CCxpher 2025-02-04 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2632930197

Mate. This is a specific bug report for NVIDIA missing textures problem.

If you're so enthusiastic about responding on how it performs on AMD cards and how to fix or optimize that, I recommend you open a proper forum thread on the subject or a bug report with AMD.

Stop hijacking NVIDIA specific bug reports on Github for unrelated discussions.

Your behavior is exactly why NVIDIA keeps those restricted internally.

Sspiffeeroo 2025-02-04 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2632958122

No you are incorrect. This is a per game tracker. I understand the OP is discussing the mesh shader problem affecting Nvidia GPUs with the buggy drivers. However, AMD users need a place to discuss the game issues affecting this game as well. AMD drivers can be buggy as well.

Hhjpaul7 2025-02-04 github

As OP, I hope @kisak-valve can kind of chime in or lock this thread temporarily:

This thread should be for all to report constructive information to help Proton developers.

Nvidia: It's a driver issue, stop spamming this thread.

AMD: Users have legit issues, let them post Proton logs or launch args.

Threads should be info gathering for developers, not a debate. If you have nothing to contribute, don't comment.

Kkisak-valve maintainer 2025-02-04 github

Hello, yes, this is the right place for all feedback related to running Final Fantasy VII Rebirth with Proton. We've structured this issue tracker that way so that we don't need to get nit picky and reject feedback when it comes to near-match topics like mods and third party builds of Proton.

Regarding video driver updates, if a video driver vendor releases a driver update and it doesn't change how a game behaves, reporting that non-functional detail doesn't help and is effectively off-topic noise.

There's no need to try to separate players based on the color of their Chocobo.

Aarnxxau 2025-02-05 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2632930197

I ran your command while the game was running, and it was operating at maximum speed (1000MHz with the 6700XT).


I've recorded two short videos comparing the frametime graphs using MangoHud on Linux and MSI Afterburner on Windows.

The Linux recording was made using GNOME Recorder.
The Windows recording looked extremely poor with the built-in recorder, so I had to use OBS instead.

On Linux, some overlay parameters are missing. I suspect this is due to the CachyOS kernel I installed for testing purposes. However, the stuttering occurs on all the kernels I’ve tested.

All Linux tests were run from the internal NVMe drive.
Previous Windows tests were done on an external SSD to check performance. The only issue with the external SSD was extreme texture pop-in, but the game ran smoothly without severe stutters.

For this Windows recording, the game is running from a small partition of my internal NVMe drive, where Linux is also installed, but formatted in NTFS so Windows can read it.

Both recordings were made with high settings, VSync enabled, and a 60 FPS cap with dynamic resolution.

The stutters persist regardless of the game’s configuration. This is just the way I’m currently playing it on Windows.

Windows:

https://github.com/user-attachments/assets/86e4ee18-33c8-4c73-a15d-08bc73d156be

Linux:

https://github.com/user-attachments/assets/5c8fa535-6716-4739-9f83-ea7b7fd5792a


Sorry for making this so long, and thanks to everyone contributing

KKurinPawpad 2025-02-05 github

Image

OS: Arch Linux
KERNEL: 6.13.1-cachyos
CPU: Ryzen 5600X
GPU: NVIDIA GeForce RTX 2060
GPU DRIVER: NVIDIA 570.86.16
Proton: Proton-Experimental [Bleeding Edge]

Just because I was curious, and my installation somehow lets me move around without issue, I decided to follow a playthrough to get through the first area a little ways (incredibly awkward with so much missing geometry), and can confirm that the bloom/lighting issue is ALSO present here as on AMD. Once you are indoors and get to a certain point, the background gets bright enough to make it difficult to see, similar to this prior post.

Jjobs-git 2025-02-05 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2633019715

Take a look at description, you are confusing people then mid way taking it back. Update your issue description to a more generic form.

SSheMelody 2025-02-05 github

As kisak just said, this is a per-game issue tracker. Both NVIDIA and AMD users have the right to type in here.

CCxpher 2025-02-05 · hidden on GitHub github

As kisak just said, this is a per-game issue tracker. Both NVIDIA and AMD users have the right to type in here.

I respectfully disagree. The report clearly states NVIDIA.

Any other argument is moot as such.

If you want to add AMD, then modify the report and indicate that.

Otherwise, file a bug report for AMD.

Tracking two completely different issues for different hardware architectures in a single report is a joke.

Zzixaphir 2025-02-05 · hidden on GitHub github

I respectfully disagree. The report clearly states NVIDIA.

Any other argument is moot as such.

If you want to add AMD, then modify the report and indicate that.

Otherwise, file a bug report for AMD.

Tracking two completely different issues for different hardware architectures in a single report is a joke.

You are literally arguing with valve staff over how they run their issue tracker and spamming people who just want to know if they can play ff7 with off topic nothings. Nobody cares about your opinion on how the issue tracker should run.

CCxpher 2025-02-06 · hidden on GitHub github

I respectfully disagree. The report clearly states NVIDIA.
Any other argument is moot as such.
If you want to add AMD, then modify the report and indicate that.
Otherwise, file a bug report for AMD.
Tracking two completely different issues for different hardware architectures in a single report is a joke.

You are literally arguing with valve staff over how they run their issue tracker and spamming people who just want to know if they can play ff7 with off topic nothings. Nobody cares about your opinion on how the issue tracker should run.

I did not argue with them.

He mentioned mods and 3rd party versions of Proton.

He did not mention AMD.

Stop framing things the way you want for your own ends.

PS : I don't care for your opinion either.

SSheMelody 2025-02-06 · hidden on GitHub github

Is there a way to kick @Cxpher out of the discussion? kisak literally stated that this is a place to post all kinds of feedback about the game yet the person keeps arguing on how Repository's owners/collaborators should run their issue tracker.

CCxpher 2025-02-06 · hidden on GitHub github

Is there a way to kick @Cxpher out of the discussion? kisak literally stated that this is a place to post all kinds of feedback about the game yet the person keeps arguing on how Repository's owners/collaborators should run their issue tracker.

I just said that if you want to post on AMD, then you should open a bug report for it. If I'm wrong, I am sure I would already be banned from this thread.

I can see however, that regardless of how right I am, there are always going to be those that insist that they have a right to do this or that.

Go ahead and knock yourselves out. I'm going to block the lot of you. That's easier and I don't have to keep coming back to read AMD responses when I'm trying to view a response to the actual bug report when it is fixed.

Nneliavaal 2025-02-06 · hidden on GitHub github

ppl are beyond stupid in this thread 🍿

SSheMelody 2025-02-06 github

"I just said that if you want to post on AMD, then you should open a bug report for it"
-> No, because kisak said that this is a per-game report. and you have no right to change the way Valve works.
I personally think Valve is doing a great job by keeping the reports per-game, imagine what a mess it would be to have a report for every problem X in the game Y. I can't even imagine it.
People are posting feedback about how the game is running on Proton for them and are being responded to. I don't see a problem with that.

Aandrew-ld 2025-02-06 github
Kkisak-valve maintainer 2025-02-06 · hidden on GitHub github

Hello @Cxpher, your choice to repeatedly spam your stance has resulted in a 24 hour timeout. You successfully communicated your opinion the first time and now it's a pile of off-topic noise.

Aarnxxau 2025-02-07 github

for red team stuttering there is a fix https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33370

I just updated to the latest Mesa Git build currently available, which includes this fix (25.1.0_devel.201497.e550a3cab04.d41d8cd-1).

However, the stuttering persists for me and feels the same as before.

Can anyone confirm if this fix has resolved the stuttering for them?

BBillli11 2025-02-07 github

@arnxxau
Using 7800xt with mesa-git, the mr do fix stuttering on my system.
Moving camera around no longer cause stuttering when setting background model to ultra or enable DRS.

With the updated caching system, there should be a new directory name radv_builtin_shaders in the games shader cache directory <steam library>/shadercache/2909400/radv_builtin_shaders

Ppotatoattack 2025-02-07 github

I can confirm the stuttering fix makes no noticeable difference on my system.
The radv_builtin_shaders directory is created in the shader cache directory.

mesa-tkg-git-25.1.0_devel.201501.08cd331cc0b
CachyOS
7900 XTX
13900k

Aandrew-ld 2025-02-07 github

@AshtenThePlatypus @arnxxau are using custom proton like proton-ge or dlss injectors like dlss enabler, optiscale ecc?

Sspiffeeroo 2025-02-08 github

The Mesa shader stutter fix is an improvement, especially in detailed areas like Kalm. Much smoother than before. However, the frame time graph is still jagged like a sawtooth.

RX 7800 XT
Arch Linux kernel 6.13.1-zen3-1-zen
Mesa 25.1.0_devel.201535.48edb9cec24
Proton Experimental 9.0 20250203
High preset
Background model detail: Ultra
In-game Framerate and Display Sync Tech: 60 fps and VRR

https://github.com/user-attachments/assets/87ff871f-2261-48df-9a5d-0d0e93e52dd9

Screenshot outside Junon while camera is turning to show frame time graph:
Image

Aandrew-ld 2025-02-09 github

@spiffeeroo the first video looks like the opening scene of the game, that part is known to lag a lot, i don't know why but even on windows and on other brands gpu's it happens, you stay there just 30 minutes.

HHansKristian-Work 2025-02-09 github

the frame time graph is still jagged like a sawtooth

The frame time graph in Mangohud isn't necessarily indicative of the actual frame pace, since that frame time graph is latched on CPU based on the present thread's progress calling vkQueuePresentKHR, not game thread. The game does struggle with frame pacing in general based on my testing (seems to be caused by IO not being as async as it could be), and when running with 144Hz VRR, the jitter is really bad, but when running with vsync on a display that is 60 Hz, the behavior improved dramatically, and it looks very smooth to me even on an aging 3950x. Using a frame limiter in-game seems to make the problem worse since there is no FIFO queue that can absorb the jitter.

I have a hunch that for the people reporting it stutters a lot and the people that say it works just fine, it might be a difference in VRR usage.

Kkhyllar 2025-02-09 github

I don't have an AMD card to see how it's running on Linux but since I'm on Nvidia I have been putting some time in on Windows.

I can say the frame times on windows are jagged as well. I also experience frequent-ish stuttering especially when I am progressing further into a zone which I think might be the game loading in new assets.

Hopefully this is helpful in determining what may be Linux performance issues versus optimization problems on FF7's end because it has been a messy release for Windows users as well.

Gguitaripod 2025-02-10 github

Fwiw, I had the dx12.2 issue and the game didn't launch on any proton version I tried. Tried changing the Wine config to Windows 11/10, but it didn't help. With the Steam launch arg -dx11 like I have in FFVIIR it gives me some other error and the game won't launch.

This was on launch. I refunded the game and haven't seen any mentions of success for gamers with similar hardware. Has anyone else?

Overall pretty disappointed Square has released yet another poor PC port (It's broken and lacking basic features on Windows too). It's inexcusable.

@arch ~ → neofetch
                   -`                    @arch
                  .o+`                   -----------
                 `ooo/                   OS: Arch Linux x86_64
                `+oooo:                  Host: Z790 AORUS ELITE AX
               `+oooooo:                 Kernel: 6.12.10-arch1-1
               -+oooooo+:                Uptime: 11 days, 4 hours, 34 mins
             `/:-:++oooo+:               Packages: 1271 (pacman)
            `/++++/+++++++:              Shell: bash 5.2.37
           `/++++++++++++++:             Resolution: 3440x1440
          `/+++ooooooooooooo/`           DE: Plasma 6.2.5
         ./ooosssso++osssssso+`          WM: kwin
        .oossssso-````/ossssss+`         Theme: Breeze-Dark [GTK2], Breeze [GTK3]
       -osssssso.      :ssssssso.        Icons: breeze-dark [GTK2/3]
      :osssssss/        osssso+++.       Terminal: ghostty
     /ossssssss/        +ssssooo/-       CPU: Intel i7-14700K (28) @ 5.500GHz
   `/ossssso+/:-        -:/+osssso+-     GPU: AMD ATI Radeon RX 7900 XT/7900 XTX/7900 GRE/7900M
  `+sso+:-`                 `.-/+oso:    GPU: Intel Raptor Lake-S GT1 [UHD Graphics 770]
 `++:.                           `-/+/   Memory: 9089MiB / 64057MiB
 .`                                 `/
Mmbriar 2025-02-10 github

@marcusziade you almost certainly got the dx12.2 not supported error because you're using amdvlk as vulkan driver instead of radv, extremely common issue on Arch that will negatively affect other games as well. Install vulkan-radeon lib32-vulkan-radeon, uninstall amdvlk lib32-amdvlk (even if you refunded the game, it will help other games as well). The game can't run with -dx11 due to it using mesh shaders.

Rredshiftltd 2025-02-12 github

Is there anything new about the Nvidia drivers on Linux? I can't play the game at all after the intro because of the missing textures. I know there's a bug report with Nvidia but none of us can see anything about it.

neofetch
/////////////                
         /////////////////////            ----------------------
      ///////*767////////////////         OS: Pop!_OS 22.04 LTS x86_64
    //////7676767676*//////////////       Kernel: 6.9.3-76060903-generic
   /////76767//7676767//////////////      Uptime: 6 days, 16 hours, 29 mins
  /////767676///*76767///////////////     Packages: 2729 (dpkg), 47 (flatpak),
 ///////767676///76767.///7676*///////    Shell: bash 5.1.16
/////////767676//76767///767676////////   Resolution: 1024x768
//////////76767676767////76767/////////   Terminal: /dev/pts/2
///////////76767676//////7676//////////   CPU: AMD Ryzen 9 5900X (24) @ 3.700GH
////////////,7676,///////767///////////   GPU: NVIDIA GeForce RTX 3060 Ti Lite
/////////////*7676///////76////////////   Memory: 56671MiB / 64224MiB
///////////////7676////////////////////
 ///////////////7676///767////////////
  //////////////////////'////////////
   //////.7676767676767676767,//////
    /////767676767676767676767/////
      ///////////////////////////
         /////////////////////
             /////////////
OOkerlund 2025-02-12 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2654998078

Sadly nothing yet. All we can really do is wait until an Nvidia driver update to hit and hope it fixes it. I am under the impression that it can't be fixed via Proton. You could maybe make a post about it on the Nvidia Linux forums and see if anyone can provide updates. https://forums.developer.nvidia.com/c/gpu-graphics/linux/148

Rredshiftltd 2025-02-13 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2654998078

I appreciate your reply and the info! I'll continue to watch this thread and also add my troubles to the Nvidia Linux forum. I see someone posted it there on 2/2.

HHansKristian-Work 2025-02-13 github

I might have found a temporary workaround for NV, but it requires some modding.

If you're brave enough, you can try building https://github.com/HansKristian-Work/pyroveil as instructed and use the config file in https://github.com/HansKristian-Work/pyroveil/tree/master/hacks/ffvii-rebirth-nvidia and it should magically work.

EDIT: Seems like dropping the file doesn't work for whatever reason, so PYROVEIL_CONFIG=/path/to/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json seems to be needed.
EDIT: You may need to delete vkd3d-proton.cache and vkd3d-proton.cache.write as well.

Bbvergnaud 2025-02-13 github

Can confirm it works on my end (at least on the opening scene). Well done you absolute genius. <3

EDIT: Seems like dropping the file doesn't work for whatever reason, so PYROVEIL_CONFIG=/path/to/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json seems to be needed.

I didn't have to explicitly set the path to the json on my end. I put it next to the exe file as documented on the readme (~/.steam/steam/steamapps/common/FINAL\ FANTASY\ VII\ REBIRTH/pyroveil.json) and used PYROVEIL=1 %command% in the game's properties in steam.
If that's relevant: GPU RTX 3070, driver 570.86.16, kernel 6.13.2-zen1

Ssteckums 2025-02-13 github

I can also report success with pyroveil by just dropping the .json directly next to ff7rebirth.exe and adding PYROVEIL=1 to the start of my launch command. I am using 570.86.15 from the cuda repository on kernel 6.13.1 on an RTX 3090.

EDIT: I am running in gamescope for HDR -- would that be why dropping the config next to the .exe works? This is my full command:

PYROVEIL=1 DXVK_HDR=1 gamemoderun gamescope -H 1440 -W 2560 -f --hdr-enabled --hdr-debug-force-output -- %command%

Zzixaphir 2025-02-13 github

Replying to [#8408 (comment)](https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2657340142)

Unfortunately, I can not report success. Here's my proton log and system info

I'm a bit of a dumbo so I probably did something wrong, but my log does report that pyroveil is loading the config.

HHansKristian-Work 2025-02-13 github

I'm a bit of a dumbo so I probably did something wrong, but my log does report that pyroveil is loading the config.

Try deleting vkd3d-proton.cache and .cache.write. It's under End/Binaries in the game install folder.

Zzixaphir 2025-02-13 github

That worked. Thanks!

DDreynbolt 2025-02-13 github

Also reporting success here with the Pyroveil workaround on Bazzite 41.

570.86.16 drivers on a 4070 Super with kernel 6.12.12-203.bazzite.fc41.x86_64

Ccyntalan 2025-02-13 github

Sadly, my attempt still failed. I deleted the cache and cache.write files but I still fail to see textures. I'm guessing that I need to have the PYROVEIL_CONFIG value stored somewhere, but where am I supposed to place it, exactly? I'm guessing the launch options?

HHansKristian-Work 2025-02-13 github

PYROVEIL=1 PYROVEIL_CONFIG=/path/to/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json %command% should work. Replace the exact path as needed. Remember to delete vkd3d-proton.cache between every failed attempt.

Ccyntalan 2025-02-13 github

Oh, pfft. I didn't actually replace the path appropriately. Derp.

That said, I did change the path accordingly, still no dice. Something I did notice is that I got no shader rebuild prompt by Steam OR in-game (which I did the first time I tried), so I'm guessing there's something still residual holding onto it?

HHansKristian-Work 2025-02-13 github

What does PROTON_LOG say if you grep for pyroveil?

Ccyntalan 2025-02-13 github

Feeling a bit dumb at the moment, but I'm getting a kickback of no such file or directory when I attempt such. Clearly I'm doing something wrong there.

UUklosk 2025-02-13 github

Many thanks @HansKristian-Work, I can confirm it is working on my side, just followed the README instructions.

OS: Arch Linux
GPU: Nvidia RTX 3080
Video driver version: 570.86.16-2
Kernel version: 6.13.2-zen1-1-zen

Image

Ccyntalan 2025-02-13 github

What does PROTON_LOG say if you grep for pyroveil?

Is there something I'm missing regarding doing the grep command listed in the readme? I tried it both leaving the appid variable in play and injecting the 2909400 id in, both with no success:

xxxxxx@xxxxxx:~$ grep "pyroveil:" ~/steam-$appid.txt
grep: /home/xxxxxx/steam-.txt: No such file or directory

xxxxxx@xxxxxx:~$ grep "pyroveil:" ~/steam-2909400.txt
grep: /home/xxxxxx/steam-2909400.txt: No such file or directory

Zzixaphir 2025-02-13 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2657714906

If you haven't, you need to add PROTON_LOG=1 to your steam command

Ccyntalan 2025-02-13 github

Ah, yeah, I was missing that.

Log grep yields:

pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3
pyroveil: Found config in /home/xxxxxxx/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json!
pyroveil: JSON parse failed: 3

Jjobs-git 2025-02-13 github

It works! RTX 4070 Super 550.144.03 6.8.0-52-generic ubuntu 24.04. This is what you will do if you are using flatpak steam, you have to install it in steam directory:

cmake . -Bbuild -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$HOME/.var/app/com.valvesoftware.Steam/.local

Image

KK1NGSSTH 2025-02-13 github

Took me some time to make it work cause i was dumb and installed Pyroveil to a different folder but it worked for me too.

OS: Arch Linux
GPU: Nvidia RTX 3070
Video driver version: 570.86.16-2
Kernel version: 6.13.2-2-cachyos

Image

00xmex 2025-02-13 github

Solution:
PYROVEIL=1 needs to be put before PYROVEIL_CONFIG=/path/to/json. My working command - after cloning pyroveil into my home, was PYROVEIL=1 PYROVEIL_CONFIG=/home/ben/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json %command%!

Also make sure that you delete vkd3d-proton.cache and vkd3d-proton.cache.write after each failed attempt!

~/pyroveil master
❯ rm -r /home/ben/.local/share/Steam/steamapps/common/FINAL\ FANTASY\ VII\ REBIRTH/End/Binaries/Win64/vkd3d-proton.cache

~/pyroveil master
❯ rm -r /home/ben/.local/share/Steam/steamapps/common/FINAL\ FANTASY\ VII\ REBIRTH/End/Binaries/Win64/vkd3d-proton.cache.write 

~/pyroveil master
❯ 

With this I, too, can confirm it to be working!
Image

It even works after the sequence and after loading into chapter 1 and its first fight!

Kernel: 6.13.2-2-cachyos
Driver: 570.86.16 (open)
GPU: 4090
Desktop: KDE 6.3
Steam: Native


Sadly I can't confirm it to be working.

Successfully build pyroveil:

❯ git clone https://github.com/HansKristian-Work/pyroveil.git
Cloning into 'pyroveil'...
remote: Enumerating objects: 36, done.
remote: Counting objects: 100% (36/36), done.
remote: Compressing objects: 100% (23/23), done.
remote: Total 36 (delta 10), reused 36 (delta 10), pack-reused 0 (from 0)
Receiving objects: 100% (36/36), 17.90 KiB | 17.90 MiB/s, done.
Resolving deltas: 100% (10/10), done.

~
❯ cd pyroveil

~/pyroveil master
❯ git submodule update --init

Submodule 'third_party/glslang' (https://github.com/KhronosGroup/glslang) registered for path 'third_party/glslang'
Submodule 'third_party/rapidjson' (https://github.com/miloyip/rapidjson) registered for path 'third_party/rapidjson'
Submodule 'third_party/shaderc' (https://github.com/google/shaderc) registered for path 'third_party/shaderc'
Submodule 'third_party/spirv-cross' (https://github.com/KhronosGroup/SPIRV-Cross) registered for path 'third_party/spirv-cross'
Submodule 'third_party/spirv-headers' (https://github.com/KhronosGroup/SPIRV-Headers) registered for path 'third_party/spirv-headers'
Submodule 'third_party/spirv-tools' (https://github.com/KhronosGroup/SPIRV-Tools) registered for path 'third_party/spirv-tools'
Submodule 'third_party/vulkan-headers' (https://github.com/KhronosGroup/Vulkan-Headers) registered for path 'third_party/vulkan-headers'
Cloning into '/home/ben/pyroveil/third_party/glslang'...
Cloning into '/home/ben/pyroveil/third_party/rapidjson'...
Cloning into '/home/ben/pyroveil/third_party/shaderc'...
Cloning into '/home/ben/pyroveil/third_party/spirv-cross'...
Cloning into '/home/ben/pyroveil/third_party/spirv-headers'...
Cloning into '/home/ben/pyroveil/third_party/spirv-tools'...
Cloning into '/home/ben/pyroveil/third_party/vulkan-headers'...
Submodule path 'third_party/glslang': checked out '633e9b6e15b7b142eb311ecc316858eab5425033'
Submodule path 'third_party/rapidjson': checked out '24b5e7a8b27f42fa16b96fc70aade9106cf7102f'
Submodule path 'third_party/shaderc': checked out 'ab0d3ed184e4919a236eecddb6514e4923a5e38f'
Submodule path 'third_party/spirv-cross': checked out '5fce5915d8f6b662501bfeb8b36f9b8918c6f813'
Submodule path 'third_party/spirv-headers': checked out 'e7294a8ebed84f8c5bd3686c68dbe12a4e65b644'
Submodule path 'third_party/spirv-tools': checked out '3364b982713a0440d1d342dd5eec65b122a61b71'
Submodule path 'third_party/vulkan-headers': checked out '234c4b7370a8ea3239a214c9e871e4b17c89f4ab'

~/pyroveil master 30s
❯ cmake . -Bbuild -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$HOME/.local
-- The CXX compiler identification is GNU 14.2.1
-- The C compiler identification is GNU 14.2.1
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
Enabling SSE3 support.
-- SPIRV-Cross: Finding Git version for SPIRV-Cross.
-- Found Git: /usr/bin/git (found version "2.48.1")
-- SPIRV-Cross: Git hash: vulkan-sdk-1.4.304.0-25-g5fce5915
-- Shaderc: build type is "Release".
-- Configuring Shaderc to avoid building tests.
-- Configuring Shaderc to build examples.
-- Configuring Shaderc to check copyrights.
-- asciidoctor was not found - no documentation will be generated
-- Found Python: /usr/bin/python3.13 (found version "3.13.1") found components: Interpreter
-- Performing Test COMPILER_SUPPORTS_FALLTHROUGH_WARNING
-- Performing Test COMPILER_SUPPORTS_FALLTHROUGH_WARNING - Success
-- Performing Test COMPILER_SUPPORTS_EXTRA_SEMI_WARNING
-- Performing Test COMPILER_SUPPORTS_EXTRA_SEMI_WARNING - Success
-- Performing Test COMPILER_SUPPORTS_PIC
-- Performing Test COMPILER_SUPPORTS_PIC - Success
-- Found Python3: /usr/bin/python3.13 (found version "3.13.1") found components: Interpreter
-- spirv-tools not linked - illegal SPIRV may be generated for HLSL
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- Configuring done (0.9s)
-- Generating done (0.1s)
-- Build files have been written to: /home/ben/pyroveil/build

~/pyroveil master
❯ ninja -C build install
ninja: Entering directory `build'
[285/286] Install the project...
-- Install configuration: "Release"
-- Installing: /home/ben/.local/lib/libVkLayer_pyroveil_64.so
-- Installing: /home/ben/.local/share/vulkan/implicit_layer.d/VkLayer_pyroveil_64.json

~/pyroveil master 45s
❯

My launch command is this (with log activated):

PYROVEIL_CONFIG=/home/ben/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json PYROVEIL=1 PROTON_LOG=1 %command

Obviously I am doing something wrong?

Edit:
I've tried to copy the pyroveil.json file next to ff7rebirth.exe. Tried with and without PYROVEIL_CONFIG= and didn't had any success. Any idea what I am doing wrong?

Using
Kernel: 6.13.2-2-cachyos
Driver: 570.86.16 (open)
GPU: 4090
Desktop: KDE 6.3
Steam: Native

Edit 2:
Even after deleting the home folder git clone and reinstalling it within FF7R folder didn't help:

❯ git clone https://github.com/HansKristian-Work/pyroveil.git

Cloning into 'pyroveil'...
remote: Enumerating objects: 36, done.
remote: Counting objects: 100% (36/36), done.
remote: Compressing objects: 100% (23/23), done.
remote: Total 36 (delta 10), reused 36 (delta 10), pack-reused 0 (from 0)
Receiving objects: 100% (36/36), 17.90 KiB | 1018.00 KiB/s, done.
Resolving deltas: 100% (10/10), done.

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH
❯ cd pyroveil

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil master
❯ git submodule update --init
Submodule 'third_party/glslang' (https://github.com/KhronosGroup/glslang) registered for path 'third_party/glslang'
Submodule 'third_party/rapidjson' (https://github.com/miloyip/rapidjson) registered for path 'third_party/rapidjson'
Submodule 'third_party/shaderc' (https://github.com/google/shaderc) registered for path 'third_party/shaderc'
Submodule 'third_party/spirv-cross' (https://github.com/KhronosGroup/SPIRV-Cross) registered for path 'third_party/spirv-cross'
Submodule 'third_party/spirv-headers' (https://github.com/KhronosGroup/SPIRV-Headers) registered for path 'third_party/spirv-headers'
Submodule 'third_party/spirv-tools' (https://github.com/KhronosGroup/SPIRV-Tools) registered for path 'third_party/spirv-tools'
Submodule 'third_party/vulkan-headers' (https://github.com/KhronosGroup/Vulkan-Headers) registered for path 'third_party/vulkan-headers'
Cloning into '/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil/third_party/glslang'...
Cloning into '/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil/third_party/rapidjson'...
Cloning into '/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil/third_party/shaderc'...
Cloning into '/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil/third_party/spirv-cross'...
Cloning into '/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil/third_party/spirv-headers'...
Cloning into '/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil/third_party/spirv-tools'...
Cloning into '/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil/third_party/vulkan-headers'...
Submodule path 'third_party/glslang': checked out '633e9b6e15b7b142eb311ecc316858eab5425033'
Submodule path 'third_party/rapidjson': checked out '24b5e7a8b27f42fa16b96fc70aade9106cf7102f'
Submodule path 'third_party/shaderc': checked out 'ab0d3ed184e4919a236eecddb6514e4923a5e38f'
Submodule path 'third_party/spirv-cross': checked out '5fce5915d8f6b662501bfeb8b36f9b8918c6f813'
Submodule path 'third_party/spirv-headers': checked out 'e7294a8ebed84f8c5bd3686c68dbe12a4e65b644'
Submodule path 'third_party/spirv-tools': checked out '3364b982713a0440d1d342dd5eec65b122a61b71'
Submodule path 'third_party/vulkan-headers': checked out '234c4b7370a8ea3239a214c9e871e4b17c89f4ab'

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil master 27s
❯ cmake . -Bbuild -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$HOME/.local
-- The CXX compiler identification is GNU 14.2.1
-- The C compiler identification is GNU 14.2.1
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
Enabling SSE3 support.
-- SPIRV-Cross: Finding Git version for SPIRV-Cross.
-- Found Git: /usr/bin/git (found version "2.48.1")
-- SPIRV-Cross: Git hash: vulkan-sdk-1.4.304.0-25-g5fce5915
-- Shaderc: build type is "Release".
-- Configuring Shaderc to avoid building tests.
-- Configuring Shaderc to build examples.
-- Configuring Shaderc to check copyrights.
-- asciidoctor was not found - no documentation will be generated
-- Found Python: /usr/bin/python3.13 (found version "3.13.1") found components: Interpreter
-- Performing Test COMPILER_SUPPORTS_FALLTHROUGH_WARNING
-- Performing Test COMPILER_SUPPORTS_FALLTHROUGH_WARNING - Success
-- Performing Test COMPILER_SUPPORTS_EXTRA_SEMI_WARNING
-- Performing Test COMPILER_SUPPORTS_EXTRA_SEMI_WARNING - Success
-- Performing Test COMPILER_SUPPORTS_PIC
-- Performing Test COMPILER_SUPPORTS_PIC - Success
-- Found Python3: /usr/bin/python3.13 (found version "3.13.1") found components: Interpreter
-- spirv-tools not linked - illegal SPIRV may be generated for HLSL
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- Configuring done (0.9s)
-- Generating done (0.1s)
-- Build files have been written to: /home/ben/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil/build

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil master
❯ ninja -C build install
ninja: Entering directory `build'
[285/286] Install the project...
-- Install configuration: "Release"
-- Installing: /home/ben/.local/lib/libVkLayer_pyroveil_64.so
-- Installing: /home/ben/.local/share/vulkan/implicit_layer.d/VkLayer_pyroveil_64.json

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil master 45s
❯ 

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil master 45s
❯ 

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil master 45s
❯ locate vkd3d-proton.cache
/home/ben/.local/share/Steam/steamapps/common/Arma Reforger/vkd3d-proton.cache
/home/ben/.local/share/Steam/steamapps/common/Arma Reforger/vkd3d-proton.cache.write
/home/ben/.local/share/Steam/steamapps/common/ChainedTogether/ChainedTogether/Binaries/Win64/vkd3d-proton.cache.write
/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/End/Binaries/Win64/vkd3d-proton.cache
/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/End/Binaries/Win64/vkd3d-proton.cache.write
/home/ben/.local/share/Steam/steamapps/common/KINGDOM HEARTS -HD 1.5+2.5 ReMIX-/vkd3d-proton.cache
/home/ben/.local/share/Steam/steamapps/common/KINGDOM HEARTS -HD 1.5+2.5 ReMIX-/vkd3d-proton.cache.write
/home/ben/.local/share/Steam/steamapps/common/Warhammer 40,000 DARKTIDE/binaries/vkd3d-proton.cache
/home/ben/.local/share/Steam/steamapps/common/Warhammer 40,000 DARKTIDE/binaries/vkd3d-proton.cache.write
/home/ben/.local/share/Steam/steamapps/common/Warhammer 40,000 DARKTIDE/shader_cache/vkd3d-proton.cache
/home/ben/.local/share/Steam/steamapps/common/sandstorm/Insurgency/Binaries/Win64/vkd3d-proton.cache
/home/ben/.local/share/Steam/steamapps/common/sandstorm/Insurgency/Binaries/Win64/vkd3d-proton.cache.write

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil master
❯ rm -r /home/ben/.local/share/Steam/steamapps/common/FINAL\ FANTASY\ VII\ REBIRTH/End/Binaries/Win64/vkd3d-proton.cache

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil master
❯ rm -r /home/ben/.local/share/Steam/steamapps/common/FINAL\ FANTASY\ VII\ REBIRTH/End/Binaries/Win64/vkd3d-proton.cache.write 

~/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil master
❯ 

Startup command was this:

PYROVEIL_CONFIG=/home/ben/.local/share/Steam/steamapps/common/FINAL\ FANTASY\ VII\ REBIRTH/pyroveil.json PYROVEIL=1 %command

Ground textures were still missing like in the old screenshots.

Edit 3:
Even with unmasking the spaces didn't help:

PYROVEIL_CONFIG=/home/ben/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json PYROVEIL=1 %command

Any suggestion will be gladly tested!

Edit 4:
Running the PYROVEIL_CONFIG path in quotation marks didn't changed anything as well.

Edit 5:
Changed from GE-Proton 9.25 to Experimental. No change as well.

Edit 6:
Turned out I still was on GE-Proton 9.24. Upgrading to 9.25 didn't fix the issue though ...

Edit 7:
If done correctly ...
This here helped me getting it to work: https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2657489797
Installed in home, set launch parameters to PYROVEIL=1 PYROVEIL_CONFIG=/home/ben/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json %command% and I finally were able to see the missing texures:
Image

NNuLLxD 2025-02-13 github

Can confirm this fix works for loading textures, however I'm softlocked after the first cutscene and can't move Zack around, which is completely unrelated. Dunno what's up with that but if anyone knows a fix please let me know. Great work @HansKristian-Work !

Image

Edit: Switching to Proton Experimental did the trick. Cheers!

Kkhyllar 2025-02-13 github

Sorry all! I am also having some issues as well and went through all the comments and am a bit confused. Are we cloning and making pyroveil inside of our rebirth directory where the .exe is located? I cloned and made it in a separate folder and copied the pyroveil.json over and it's not working. Thank you for any help on this!

NNuLLxD 2025-02-13 github

@khyllar

Sorry all! I am also having some issues as well and went through all the comments and am a bit confused. Are we cloning and making pyroveil inside of our rebirth directory where the .exe is located? I cloned and made it in a separate folder and copied the pyroveil.json over and it's not working. Thank you for any help on this!

These were my steps to get it to work;

  • Clone and build in ~/Downloads. Make sure if you are using the flatpak version of steam you use the appropriate install prefix with cmake.
  • Copy the pyroveil.json from Downloads/pyroveil/hacks/ffvii-rebirth-nvidia and add it to ~/Final Fantasy VII Rebirth (the same directory as ff7rebirth.exe).
  • Set your Final Fantasy VII REBIRTH launch options to PYROVEIL=1 %command%

With this you should be all set. If there are any issues you can add PROTON_LOG=1 to your launch command and grep the log to see if pyroveil is loaded properly.

Kkhyllar 2025-02-13 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2657931953

Thank you! Unfortunately it's still not working. I have steam installed through pacman. Installed Pyroveil to an Apps folder I have in my home directory.

I have tried both pasting the .json file in the .exe directory and pointing to it in my launch options as well as pointing to the .json in my pyroveil folder within my Apps directory. I have also deleted the vkd3d files as directed above.

here are my logs when I grep pyroveil. Which for me it was a steam-2909400.log file not .txt in case anyone runs into the same.

pyroveil: Found config in /home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json!
pyroveil: Adding match for spirvExecutionModel: 5365 (MeshEXT).
pyroveil: Adding GLSL roundtrip via SPIRV-Cross for match.
pyroveil: Found config in /home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json!
pyroveil: Adding match for spirvExecutionModel: 5365 (MeshEXT).
pyroveil: Adding GLSL roundtrip via SPIRV-Cross for match.
pyroveil: Found config in /home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json!
pyroveil: Adding match for spirvExecutionModel: 5365 (MeshEXT).
pyroveil: Adding GLSL roundtrip via SPIRV-Cross for match.
pyroveil: Found config in /home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json!
pyroveil: Adding match for spirvExecutionModel: 5365 (MeshEXT).
pyroveil: Adding GLSL roundtrip via SPIRV-Cross for match.
pyroveil: Found config in /home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json!
pyroveil: Adding match for spirvExecutionModel: 5365 (MeshEXT).
pyroveil: Adding GLSL roundtrip via SPIRV-Cross for match.
pyroveil: Found config in /home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json!
pyroveil: Adding match for spirvExecutionModel: 5365 (MeshEXT).
pyroveil: Adding GLSL roundtrip via SPIRV-Cross for match.
pyroveil: Found config in /home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json!
pyroveil: Adding match for spirvExecutionModel: 5365 (MeshEXT).
pyroveil: Adding GLSL roundtrip via SPIRV-Cross for match.
pyroveil: Found config in /home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json!
pyroveil: Adding match for spirvExecutionModel: 5365 (MeshEXT).
pyroveil: Adding GLSL roundtrip via SPIRV-Cross for match.
pyroveil: Found config in /home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json!
pyroveil: Adding match for spirvExecutionModel: 5365 (MeshEXT).
pyroveil: Adding GLSL roundtrip via SPIRV-Cross for match.

and these are my launch options currently

gamemoderun mangohud PROTON_LOG=1 PYROVEIL=1 PYROVEIL_CONFIG=/home/xxxx/Apps/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json %command%

EDIT: Dumb guy here sorry.... It's working now everything I posted here was correct but I hadn't deleted the new vkd3d files since I was doing the wrong things. Going back and deleting them again has pyroveil working now. I'll keep this up in case any of the details are helpful to someone else. Thank you @NuLLxD !

And thank you @HansKristian-Work you absolute hero. I don't know how you did it but thank you!

Rredshiftltd 2025-02-14 github

PYROVEIL=1 PYROVEIL_CONFIG=/path/to/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json %command% should work. Replace the exact path as needed. Remember to delete vkd3d-proton.cache between every failed attempt.

Really appreciate your work, thank you! Can confirm the game is playable on my RTX 3060Ti.

Now to set about getting it to run well....

SSloogs 2025-02-14 github

It works! Thank you so much. :)

My specs:

Fedora 41
Kernel 6.12.11-200
CPU: Ryzen 9 7950X
GPU: Nvidia GeForce GTX 3060 Ti
Driver version: 570.86.16
Proton Experimental [bleeding-edge]

Eevrowe 2025-02-14 github

@HansKristian-Work Amazing, it works exactly as advertised! Thank you for continuing to put your time into this one, I for one really appreciate it.

OS: Bazzite F41
GPU: RTX 3080 Ti
Kernel: 6.12.12-203
Driver Version: 570.86.16
Proton Experimental

Image

Jjobs-git 2025-02-14 github

steam-2909400.log

OS: Ubuntu 24.04
GPU: RTX 4070 Super
Driver: 550
Proton: Experimental

Encountering graphical glitch after skipping scenes:

Image

Ttaron-sys 2025-02-14 github

I can confirm that the workaround works.
I've fired them somewhere orderly, created a Build Folder for it under ./home . Then copied the .json from the hacks folder to beside the .exe of the game. Also added the Launch Options PYROVEIL=1 %command%

Tried it on these two systems and it worked, I'm using GE-Proton 9-25 on both.

First System:

  • Zbook Fury 16 G10
  • Fedora 41
  • RTX2000 Ada Laptop GPU
  • Driver: 570.86.16-5

Second System:

  • Custom PC
  • Fedora 41
  • RTX4060ti 16GB
  • Driver: 570.86.16-5

Edit: If you run into stutter while moving the camera: switching from VRR to Vsync fixed it for me. The Framerate is very stable.

Ccyntalan 2025-02-14 github

I think I discovered my problem. The json I grabbed was not the same file as the one I had locally. Copying the local json seemed to do the trick.

Awesome work, @HansKristian-Work !

Rredshiftltd 2025-02-14 github

So now I can move around in the game but it runs very poorly even on lowest settings, maybe 15-20 FPS even for the cinematics. I really appreciate @HansKristian-Work for stepping in and helping out but I don't really want to play if it's going to be like moving through sludge.

I set everything to Low, running at 1920x1080 on this hardware:

CPU: AMD Ryzen 9 5900X (24) @ 3.700GH
GPU: NVIDIA GeForce RTX 3060 Ti Lite
RAM: 64GB

Given that the low spec for the game is an RTX 2060 I'd hoped it would be playable on Low via Proton but I'm getting nowhere near the FPS that Windows users are reporting. Not sure what to do though, maybe set up GPU passthrough to a Windows VM? Or wait for a driver update?

For comparison, Returnal runs great on this system on High settings. I really don't think it's that my hardware is lacking.

GGalacticDeep 2025-02-14 github

Intel Arc B580

Game does not launch. Error message comes up, "DX12 is not supported on your system. Try running without the -dx12 or -d212 command line argument."

System Information

GPU: Intel Arc B580 
Video driver version: Mesa 25.1.0-devel (git-6652eb0ec3)
Kernel version: 6.13.2-zen1-1-zen 
Link to full system information report: https://gist.github.com/GalacticDeep/0c8949dc48d0e224ec52e5535f1209c3
Proton version: Experimental
OOpenLogicEFI 2025-02-14 github

@HansKristian-Work Thank you so much!!! Working great so far!!!

OS: Mint 22 Cinnamon
CPU: Ryzen 7 7435HS
GPU: RTX 4070 mobile max-q
Kernel: 6.8.0-53-generic
Driver Version: 570.86.16
Proton Experimental

Image

Ccyberiacafe 2025-02-14 github

Forgot to say thanks for the pyroveil fix! been playing all day!! one thing i suggest if anyone is having crashes still with it, limit your monitor to 60hz!!

OS:Debian 12
CPU: AMD Ryzen 7 2700x
GPU:RTX 3060ti
Kernel: 6.1.0-30-amd64
NVIDIA Driver Version: 570.86.16
Proton Experimental

Image

Ppotatoattack 2025-02-15 github

I just want to further document my stuttering experience with this game.

  • OS: CachyOS
  • CPU: Intel i9-13900k
  • GPU: RX 7900 XTX
  • Kernel: Linux 6.13.2-2-cachyos
  • DE: KDE Plasma 6.3.0
  • mesa: mesa-tkg-git 25.1.0_devel.201804.a9b6a54a8cc-1
  • Proton versions tested:
    • Steam's default (for this game)
    • experimental
    • GE
    • proton-cachyos

As suggested by @HansKristian-Work It could be a VRR issue, the game and monitor's refresh rate don't seem that locked in.
Usually when games have VRR stutters the monitors refresh briefly goes to max (240 hz in this case) which does not happen here.

The monitors refresh rate is in the top right.

https://github.com/user-attachments/assets/0f03d9bb-1394-492a-82c7-04fd58553a0a

The issue is the same running without any launch options and with my usual launch options:
LD_PRELOAD="" game-performance gamescope -w 3840 -h 2160 -r 240 -f --hdr-enabled --adaptive-sync -- %command%

It's also present at all in game quality settings but does seem to get a little better on Low.

WWhitleyStriber 2025-02-15 github

Image

3080Ti. Thank you very much.

Aarnxxau 2025-02-15 github

@arnxxau Using 7800xt with mesa-git, the mr do fix stuttering on my system. Moving camera around no longer cause stuttering when setting background model to ultra or enable DRS.

With the updated caching system, there should be a new directory name radv_builtin_shaders in the games shader cache directory <steam library>/shadercache/2909400/radv_builtin_shaders

I do have the new folder in shadercache. I also cleaned the prefix and the shadercache folder before testing.

@AshtenThePlatypus @arnxxau are using custom proton like proton-ge or dlss injectors like dlss enabler, optiscale ecc?

Yes. I'm using protonge and optiscaler, but I also tried proton experimental and completely disabling optiscaler, and that didn’t change anything.

Aarnxxau 2025-02-15 github

Replying to [#8408](https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2660607517)

I tried disabling VRR everywhere, but I still get the same stutters.

If VRR was the issue, then why does turning it off in-game, in the desktop manager, and in the display settings not fix it?

RRoboRev 2025-02-15 github

Also reporting success here with the Pyroveil workaround on Bazzite 41.

570.86.16 drivers on a 4070 Super with kernel 6.12.12-203.bazzite.fc41.x86_64

I am still new to linux (using Bazzite), can I ask for your help to get this working?

I followed the instructions on the Pyroveil Git, but I get met with this:
bazzite@bazzite:~/pyroveil$ cmake . -Bbuild -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$HOME/.local CMake Error: CMake was unable to find a build program corresponding to "Ninja". CMAKE_MAKE_PROGRAM is not set. You probably need to select a different build tool. CMake Error: CMAKE_CXX_COMPILER not set, after EnableLanguage CMake Error: CMAKE_C_COMPILER not set, after EnableLanguage -- Configuring incomplete, errors occurred!

I tried installing Ninja but I don't think I am doing it right... I really don't know what I am doing... I just wanna play FFVIIRebirth

Mm4gpi 2025-02-15 github

Anyone got any advice? I can't get the game to start. I installed and setup pyroveil and it appears to find the config file (see logs) so I'm not sure what the issue is!

steam-2909400.log

Linux Mint 22 Cinnamon
6.11.0-17-generic
RTX 3050
570.86.16
Proton experimental

SSloogs 2025-02-16 github

I'm definitely noticing a lot of stuttering in cutscenes while using pyroveil on Nvidia, although gameplay seems fairly smooth so far. This occurs even on Low settings. I'm running the game off an M.2 drive as well.

I also noticed some flickering behaviour with some of the lighting, along with light sources popping in and out. Has anyone else experienced this?

Just wondering if a anyone has advice or workarounds.

Fedora 41
Kernel 6.12.11-200
CPU: Ryzen 9 7950X
GPU: Nvidia GeForce GTX 3060 Ti
Driver version: 570.86.16
Proton Experimental [bleeding-edge]

Sspiffeeroo 2025-02-16 github

I tried disabling VRR everywhere, but I still get the same stutters.

If VRR was the issue, then why does turning it off in-game, in the desktop manager, and in the display settings not fix it?

Set the monitor refresh rate to your frame rate cap and use V-sync. So set a 60 fps limit within the game's setting, set the monitor refresh rate to 60 Hz. Do not use an external frame limiter.

If your in-game fps drops below the frame rate cap, then the frame time becomes jittery again. Probably only set 120 Hz refresh rate on the monitor if you can guarantee 120 fps minimum.

I have a 240 Hz monitor, but only setting the monitor refresh rate to 60 Hz mostly fixes the frame time graph with a 60 fps in-game cap. Changing the monitor refresh rate to 120 Hz or 240 Hz refresh rate is still jittery, even though conventional Internet wisdom says multiples of 60 Hz refresh rates should be okay. This game has a quirky engine.

This screenshot is when I spin the camera around in Kalm.

Image

Aatobenzo 2025-02-16 github

CPU: AMD Ryzen 7 7840U
GPU: Nvidia GeForce GTX 3060 Ti
nvidia 570.86.16-5 (Arch)
Proton Experimental [bleeding-edge]

pyroveil lets me play the game, but it only runs at about 22fps @ 1080p. Not sure if that's an expected result of the workaround or if there's something else going on.

If I run with the same graphics settings on a Windows VM with GPU passthrough I can get 60fps fairly consistently.

Bbleuthoot-sven 2025-02-16 github

Some installation instructions for anyone running on Bazzite (or similar). You will need to run this in a distrobox container, preferably a fedora one

  1. Create a Fedora distrobox if you don't already have one: distrobox create fedora (You can also make one using BoxBuddy, which should already be installed on Bazzite. Make sure you select the registry.fedoraproject.org/fedora-toolbox image)
  2. Enter the distrobox using distrobox enter fedora
  3. Install the required dependencies: sudo dnf install make automake gcc gcc-c++ kernel-devel ninja-build
  4. Clone the repository: git clone https://github.com/HansKristian-Work/pyroveil.git
  5. Enter the folder using cd pyroveil
  6. Run git submodule update --init to download the extra submodules
  7. Run CMake cmake . -Bbuild -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$HOME/.local
  8. Run ninja to install the library: ninja -C build install
  9. Copy the pyrofveil.json file from hacks/ffvii-rebirth-nvidia to the installation directory (Steam > Final Fantasy VII REBIRTH > Right click the game > Browse Local Files)
  10. Delete all the the cache.write files from the End/Binaries/Win64/ where FF7Rebirth is installed
  11. Edit the launch options of FF7Rebirth in Steam to the following: PYROVEIL=1 %command%
  12. Launch the game

@RoboRev Hopefully this can help

CCxpher 2025-02-16 github

I think most people who use NVIDIA cards and have problems with getting it fixed might miss the step of deleting the vkd3d proton cache.

Don't make a misatake with this command below.

rm -rf $HOME/steaminstallfolder/steamapps/common/FINAL FANTASY VII REBIRTH/End/Binaries/Win64/vkd3d-proton.cache

If you're unsure, change directory in the folder first before deleting the cache file.

Then start FFVII and you're good to go.

Mmlbarker 2025-02-16 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2661058683

Make sure you download and install using this: sudo rpm-ostree install ninja-build

CCxpher 2025-02-16 github

I'd add that it works with ultrawide fix as well. So for those of you playing on like 7680x2160 resolution for example, this combination works well.

Fix for shaders + ultrawide fix.

TTikilou 2025-02-16 github

CPU: AMD Ryzen 7 7840U GPU: Nvidia GeForce GTX 3060 Ti nvidia 570.86.16-5 (Arch) Proton Experimental [bleeding-edge]

pyroveil lets me play the game, but it only runs at about 22fps @ 1080p. Not sure if that's an expected result of the workaround or if there's something else going on.

If I run with the same graphics settings on a Windows VM with GPU passthrough I can get 60fps fairly consistently.

Same here with GTX 1650.

PPaddlefruit 2025-02-16 github

Well @HansKristian-Work's solution works for getting the textures in loading, frequently a lot of the characters or environment geometry will explode in a bunch of tris, haha. Anyone know a workaround?

OS: Linux Mint 21.3 Cinnamon
GPU: RTX 3080 Mobile

Image

RRoboRev 2025-02-17 github

3. Install the required dependencies: sudo dnf install make automake gcc gcc-c++ kernel-devel ninja

@bleuthoot-sven First thank you so much for the help here.

I am getting an error at step 3 saying that it is unable to find a match : ninja

I think this is the same issue I was having before - Ninja not found

RRoboRev 2025-02-17 github

@bleuthoot-sven Okay so I THINK I was able to install ninja properly by having it install ninja-build instead of ninja in the command line. (from what I understand they are the same thing...)

But now I am having an issue where cloning the git doesn't clone everything properly. when I try to run cmake, I get a bunch of errors all saying the same thing:

"PATH TO FOLDER" does not contain a CMakeLists.txt file.

When I look in the Pyroveil directory, this is true. Looking in the Git repository, there seems to be an insane amount of files not being copied over, including said .txt files. Even when I manually download the zip file from github, those files are still missing. Really not sure what is going on here

Bbleuthoot-sven 2025-02-17 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2663863819

@RoboRev

Whoops, I forgot to add a few steps. Added them.

FYI: all of the steps, except the ones installing the required packages/setting up the distrobox container/deleting the cache, are from HansKristian-Work github page.

RRoboRev 2025-02-17 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2663880244

AMAZING!! So before I read this update, I manually downloaded and extracted each subfolder into their respective proper folders lol. Followed everything else and BAM! WE ARE UP AND RUNNING!! THANK YOU SO MUCH FOR THE HELP, YOU ARE A HERO

TTheCliffRacer 2025-02-17 github

Thanks for the fix but the game randomly crash during fights with the following error:

NVRM: Xid (PCI:0000:01:0 0): 31, pid=240017, name=ff7rebirth_.exe, Ch 00000036, intr 00000000. MMU Fault: ENGINE GRAP HICS GPC4 GPCCLIENT_GCC faulted @ 0xfffffdf0_0033f000. Fault is of type FAULT_PDE ACCESS_TYP E_VIRT_READ

Is it somehow related to the workaround mentioned above?

Aatobenzo 2025-02-17 github

CPU: AMD Ryzen 7 7840U GPU: Nvidia GeForce GTX 3060 Ti nvidia 570.86.16-5 (Arch) Proton Experimental [bleeding-edge]

pyroveil lets me play the game, but it only runs at about 22fps @ 1080p. Not sure if that's an expected result of the workaround or if there's something else going on.

If I run with the same graphics settings on a Windows VM with GPU passthrough I can get 60fps fairly consistently.

I found the solution for this: for whatever reason, if I set my Wayland plasma desktop resolution to 1440p and play at that resolution then the game runs at ~60fps, but if the desktop resolution is 4k and I set the in-game resolution to 1080p or 1440p it runs 20-30fps. I don't really understand why that would be the case, maybe something to do with using an eGPU with limited bandwidth.

Bbernardodpc 2025-02-18 github

@Paddlefruit I'm experiencing the same thing after playing for a while. I have to quit to tittle screen and reload the game in order to solve all those black polygons.

RTX 3050 Mobile
ubuntu 24.04

CCeeJay654 2025-02-19 github

Doesn't work for me. Seems I don´t get it. I tried everything but still no textures. Seems I build it wrong, but I really don´t know. I wish there was a video tutorial. Messing with it for 4 days now.

Games is Epic Games Version on Heroic Games Launcher.

System

Garuda Linux (Arch) on Kernel 6.12.1-1-cachyos-bore
AMD Ryzen 5 5600X
Nvdia RTX 4060 Ti
32 GB Ram

TTheCliffRacer 2025-02-19 github

Doesn't work for me. Seems I don´t get it. I tried everything but still no textures. Seems I build it wrong, but I really don´t know. I wish there was a video tutorial. Messing with it for 4 days now.

System

Garuda Linux (Arch) on Kernel 6.12.1-1-cachyos-bore AMD Ryzen 5 5600X Nvdia RTX 4060 Ti 32 GB Ram

@CeeJay654 could you please check proton logs to see if PYROVEIL was loaded correctly?

CCeeJay654 2025-02-19 github

Not much time right now. A bit busy.

Where is the log in

/.steam/steam/logs/

TTheCliffRacer 2025-02-19 github

Not much time right now. A bit busy.

Where is the log in

/.steam/steam/logs/

@CeeJay654 Proton logs should be enabled first with the following command: PROTON_LOG=1. Then try to run the game. The output location should be: $HOME/steam-<STEAM_GAME_ID>.log, for Steam in flatpack: $HOME/.var/app/com.valvesoftware.Steam/steam-<STEAM_GAME_ID>.log. Proton output location can be overridden with: PROTON_LOG_DIR

Aabelbmx 2025-02-19 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2610532289

Does it work with a GTX? Is this only for Linux or does it also work for Windows?

Mmbriar 2025-02-19 github

@abelbmx no it doesn't work, while the override maybe tricks the game into starting, it crashes later while compiling shaders. There is no workaround if it requires mesh shaders.

WWhitleyStriber 2025-02-19 github

Thanks for the fix but the game randomly crash during fights with the following error:

NVRM: Xid (PCI:0000:01:0 0): 31, pid=240017, name=ff7rebirth_.exe, Ch 00000036, intr 00000000. MMU Fault: ENGINE GRAP HICS GPC4 GPCCLIENT_GCC faulted @ 0xfffffdf0_0033f000. Fault is of type FAULT_PDE ACCESS_TYP E_VIRT_READ

Is it somehow related to the workaround mentioned above?

I too am also experiencing this for the 1st time in Chapter 3. I have to save before each fight

Linux

nvidia-bug-report.log.gz

6.13.3-arch1-1 #1 SMP PREEMPT_DYNAMIC Mon, 17 Feb 2025 17:42:11 +0000 x86_64 GNU/Linux

DDragonslayer12302 2025-02-19 github

Anyone know how to fix this? Insane stuttering ONLY when panning camera, using the workaround to make textures work with NV gpu. Fedora 41, proton bleeding edge(which ended up stuttering the least). Shaders are already compiled.
https://github.com/user-attachments/assets/98ebf43b-3d5c-46c5-b327-4592c82140d8

https://www.youtube.com/watch?v=U8xTyT5eY-0 just watched this video on a 4070Ti and it has nearly zero stutters. What is wrong with my config??

Aalexiscirmi 2025-02-20 github

Anyone know how to fix this? Insane stuttering ONLY when panning camera, using the workaround to make textures work with NV gpu. Fedora 41, proton bleeding edge(which ended up stuttering the least). Shaders are already compiled. https://github.com/user-attachments/assets/98ebf43b-3d5c-46c5-b327-4592c82140d8

https://www.youtube.com/watch?v=U8xTyT5eY-0 just watched this video on a 4070Ti and it has nearly zero stutters. What is wrong with my config??

Same issue here, RTX 4060 on Fedora 41, 565.77 driver. Tried Proton Bleeding Edge and GE 9-25.

Jjorgevp 2025-02-20 github

Running cmake produces a warning:

-- spirv-tools not linked - illegal SPIRV may be generated for HLSL

Will this affect whether the hotfix works? Still missing textures, although the ninja build installed its libraries.

My proton log does not show logs from pyroveil, I've cleared the cache each attempt and my launch options are

PROTON_LOG=1 PROTON_LOG_DIR=/home/kapao/ PYROVEIL=1 PYROVEIL_CONFIG=/home/kapao/Downloads/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json %command%

I installed steam through snap if that helps, too

RRoyShapiro 2025-02-20 github

@Dragonslayer12302 , @alexiscirmi While it might not help everybody, what helped me against the camera panning stuttering on an 8 Gb VRAM GPU ( GeForce RTX 3060 Ti ) was to lower the Background Model Detail (or what is it called) as well as the Texture Resolution all the way down to Low (basically as low as it can go). This is, as far as I have noticed, likely because the game constantly tries to render the objects in and out of view as you rotate the camera, and the more detailed the objects are, the more gittery it becomes. This game is also an overwhelming VRAM hog: Even setting the Textures to Medium on such a seemingly capable GPU causes it to run out of memory and begin stuttering in a lot of locations in mere 10-20 minutes. I've measured VRAM usage in Mangohud, but the symptom is unmistakable, as soon as it goes over 7.9 Gb, stuttering ensues.

BBitBlitObviMormon 2025-02-20 github

Pyroveil mostly works, but after Nibelheim I get a similar problem as @Paddlefruit. After a few minutes of playing the tris on the characters and terrain explode and flash rapidly, covering the screen. For me it doesn't go away until I delete the vkd3d proton cache again and restart the game. The issue always returns after playing for a few minutes. Any idea what might be happening here or if there is a workaround?

Image

Image

Specs
OS: Linux Mint 22
KERNEL: 6.8.0-53-generic
CPU: Intel Core i7-14700F
GPU: NVIDIA GeForce RTX 4070
GPU DRIVER: NVIDIA 550.120
RAM: 32 GB

Aalexiscirmi 2025-02-20 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2670915830

I guess this is driver related. I've seen an RX 6600 not having this issue on Linux. Yes, less fps but not this stuttering. Let's hope it'll be fixed on the next Nvidia driver version.

DDragonslayer12302 2025-02-20 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2670915830

I have already tried setting everything to the lowest settings even at 720P with 33% DLSS, it does not help.

Regardless, I know the 24GB VRAM on my card is not an issue.

BBitBlitObviMormon 2025-02-20 github

For those who are seeing the exploding tris, I updated my graphics driver from 550 to 570 and the issue never returned. I am not sure if this is the fix or if the fix is a side effect from forcing a recompilation of all of the game's shaders.

Jjobs-git 2025-02-21 github

For those who are seeing the exploding tris, I updated my graphics driver from 550 to 570 and the issue never returned. I am not sure if this is the fix or if the fix is a side effect from forcing a recompilation of all of the game's shaders.

The issue with 570 is that it causes alt-tab problems on majority of games. 550 is the most stable

NNycz-lab 2025-02-21 github

@jobs-git afaik that is caused by VRR, you can just deactivate it for now then you wont get the alt-tab blackscreen

Jjobs-git 2025-02-21 github

@jobs-git afaik that is caused by VRR, you can just deactivate it for now then you wont get the alt-tab blackscreen

That is weird because some games also need VRR to have stable frametimes (i.g. rebirth)

Ppavethem 2025-02-21 github

I just got the exploding triangles issue using pyroveil yesterday as well. For me it started in the Mythril Mines, weird how it seems to pop up at different spots for everybody. Deleting the vkd3d-proton.cache does not seem to help either.
The flickering seems to not happen in Proton 9.0-4, however this version freezes constantly for me for whatever reason.

CPU: Intel I7-7700K
GPU: Nvidia GeForce GTX 3080 Ti
Nvidia 570.86.16-8
Arch Linux
Proton Experimental

GGuillaume-Gillon 2025-02-22 github

I would like to share my experience about pyroveil.

First of all, thanks a lot to @HansKristian-Work for this amazing work!!

Steam was installed on my laptop by a snap package and I suppose the mod isn't working in this case. I had to remove steam and reinstall the .deb package downloaded from steam website to see that software directories are different with this way to install it. After this, following the readme steps was making the job and the solution worked very well.

So for anyone who have issues with pyroveil, maybe this solution could help.

And now, let's play!! 😁

Ubuntu 24.04 - Wayland
Nvidia 550 drivers

Bbernardodpc 2025-02-23 github

@pavethem

weird how it seems to pop up at different spots for everybody

yes, it's weird indeed... it was very casual for me on chapter 2 and 3 but after mythril mine it comes now very often. exiting to title screen and reloading the game makes it go away for me but it's very annoying since it's happening every 2 minutes.

RTX 3050 Mobile
Nvidia 550 drivers
Ubuntu 24.04 - wayland

Yyscialom 2025-02-23 github

Reporting success with NVIDIA + pyroveil.

System

  • CPU: AMD Ryzen 9 5900X (24) @ 4.951GHz
  • GPU: NVIDIA GeForce RTX 3060 Lite Hash Rate
  • Drivers: 560.35.03 | CUDA Version: 12.6
  • Memory: 32G DDR4
  • OS: Ubuntu 24.10 x86_64
  • Kernel: 6.11.0-18-generic
  • DE: GNOME 47.0 + Wayland
  • WM: Mutter

Configuration of pyroveil

  • clone & build
  • Drop pyroveil.json somewhere locally
  • Configure Steam to run the game with:
    -- PYROVEIL=1 PYROVEIL_CONFIG="/chosen/path/pyroveil.json" %command%
    -- Proton experimental

Success

Game works, no tris explosion, no crashes.
Performance: 45-60fps @ 4k with some annoying stutters

Ddragaoazuljr 2025-02-24 github

PSA: Issues Running Heroic (Flatpak) with Pyroveil

Issue

If you're trying to use Pyroveil with Heroic (Flatpak), you may run into issues where Pyroveil doesn’t seem to execute at all. The game launches, but Pyroveil logs do not appear in the Proton log, indicating that it is not being loaded.

Troubleshooting Steps Taken

  • Set the required environment variables in Heroic.
  • Cleared the VK3D cache.
  • Checked Proton logs (PROTON_LOG=1), but there was no mention of Pyroveil.
  • Verified that Heroic (Flatpak) has limited access to certain directories (~/Games/Heroic, /mnt, /run/media, etc.).
  • Used Flatseal to grant additional permissions, including access to:
    • The entire pyroveil folder.
    • ~/.local/share/vulkan (where Pyroveil installs some files).
  • Attempted to run the Pyroveil build command specific for Flatpak Steam, but pointed it to the .var directory used by Heroic. This also did not resolve the issue.
  • Despite these changes, Pyroveil still did not work.

Solution

Switching from the Flatpak version of Heroic to the AppImage version resolves the issue. With the AppImage version:

  • The game renders objects successfully.
  • Pyroveil logs correctly appear in the Proton log.

This suggests that either:

  1. Flatpak restrictions are preventing Pyroveil from accessing the necessary files.
  2. Some additional directories still need to be made accessible in Flatpak for Pyroveil to work.

Note: Since Lutris is also available as a Flatpak, it is possible that the same issue could occur when using Pyroveil with Lutris (Flatpak). If anyone has tested this, it would be useful to confirm whether the same workaround (using the non-Flatpak version) is necessary.


Additional Issue: Freezing

Even after resolving the render issue, I’m running into frequent freezing in the game. The game crashes within minutes, sometimes even during cutscenes.

Steps Tried

  • Enabled VSync both in-game and at the monitor level (no success).
  • Tried different Proton versions (Proton Experimental, Proton GE), but the issue persists.
  • Running the game through Gamescope with the following command:
  gamescope -W 2560 -H 1440 -r 60  -- %command%

Someone suggested setting the v-sync flag on gamescope bu the only potentially relevant flag I found for VSync in Gamescope is -r (refresh rate), which is set to 60, matching my monitor.

System Specs

  • CPU: Intel Xeon E5-2680 v4
  • GPU: NVIDIA GeForce RTX 3060 Lite Hash Rate
  • Drivers: 570
  • Memory: 16GB DDR4
  • OS: Bazzite 41 Gnome Nvidia

If anyone has suggestions for fixing the freezing issue, I’d appreciate any insights!

Jjorgevp 2025-02-24 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2676413599

I switched to the deb install from the snap package. This fixed it.
@HansKristian-Work Pyroveil likely needs a different install location with snap, similar to how flatpak's install dir is different. I tried dropping the library to several other lib folders within the snap directory to no avail, I'm not too well versed how steam differs when using different installs.

Hhivehivemind 2025-02-25 github

RX 7900 XT
5800X3D

It's a generally great experience based on latest VKD3D and RADV git. There is a crash on the exact same spot every time on chapter 12, in which driving the boat through the northern part of open sea crashes the game. More specifically, the game flashes white, freezes and crashes the video driver.
Not a VRAM leak, although it seems like a shader related memory violation.

It is also reproducible on stable Mesa, older Proton 9.0.5 and VKD3D-Proton with Rebirth's shader quirks (before this commit).

Attaching my whole FINAL FANTASY VII REBIRTH folder that goes in ../My Documents/My Games/. It includes my save file for testing. You need to drive the boat around the gray area just north of this save's shore.

Oddly enough, driving through the whole rest of the map it works just fine, just like playing the rest of the game overall (~110 hours in)

FINAL FANTASY VII REBIRTH.zip

Kkisak-valve maintainer 2025-02-25 github

Hello @hivehivemind, please add PROTON_LOG=1 %command% to the game's launch options, reproduce the crash and attach the generated $HOME/steam-$APPID.log to this issue report as a file. (Proton logs compress well if needed.) Also, please copy the contents of Steam Runtime Diagnostics from Steam (Steam -> Help -> Steam Runtime Diagnostics) and put it in a gist, then include a link to the gist in this issue report.

Kkisak-valve maintainer 2025-02-25 github

Hello @IanLugli, we can not help with pirated games.

HHansKristian-Work 2025-02-26 github

There is a crash on the exact same spot every time on chapter 12, in which driving the boat through the northern part of open sea crashes the game. More specifically, the game flashes white, freezes and crashes the video driver.

Do you have any information on exactly where that spot is? A screenshot showing where on the map it is should be enough to repro. The open sea isn't exactly small, and I didn't run into any issues when I played through that section.

You need to drive the boat around the gray area just north of this save's shore.

I guess this might be enough to repro.

Hhivehivemind 2025-02-27 github

Do you have any information on exactly where that spot is? A screenshot showing where on the map it is should be enough to repro. The open sea isn't exactly small, and I didn't run into any issues when I played through that section.

It is right around this area, more specifically:

Image

NNycz-lab 2025-02-27 · hidden on GitHub github

nvidia released driver 570.124.04 an hour ago
did this driver change anything about how the game runs?
Changelog doesnt address these issues at all: https://www.nvidia.com/en-us/drivers/details/241089/

Ddragaoazuljr 2025-02-27 github

Just an update on the freezing issue: when not using gamescope the problem seems to occur less frequently.

HHansKristian-Work 2025-02-28 github

@hivehivemind We found another game sync bug. A workaround was pushed to vkd3d-proton master. https://github.com/HansKristian-Work/vkd3d-proton/pull/2362 Hopefully that fixes the hang.

HHansKristian-Work 2025-03-01 github

The new Vulkan beta driver from NVIDIA 570.123.01 (https://developer.nvidia.com/vulkan-driver) was reported by them to resolve the bug that pyroveil hacked around. I ran around in Nibelheim to test and it seems to work as expected now.

Jjobs-git 2025-03-01 github

The new Vulkan beta driver from NVIDIA 570.123.01 (https://developer.nvidia.com/vulkan-driver) was reported by them to resolve the bug that pyroveil hacked around. I ran around in Nibelheim to test and it seems to work as expected now.

tnx, now the game freezes and are stuck on Ubuntu 24.04

Mmrhanman 2025-03-01 github

The new Vulkan beta driver from NVIDIA 570.123.01 (https://developer.nvidia.com/vulkan-driver) was reported by them to resolve the bug that pyroveil hacked around. I ran around in Nibelheim to test and it seems to work as expected now.

I just installed Nvidia driver version 570.124.04 on Arch with the 6.12.17 lqx kernel and without pyroveil, it still does not render textures. Am I misunderstanding that this should include the new vulkan-driver?

KKurinPawpad 2025-03-01 github

Can confirm arch nvidia driver 570.124.04 on 6.13.3-1-cachyos still does not render mesh textures, even after removing the vkd3d-proton.cache(and .cache.write) files and forcing shader rebuild.

FFlat 2025-03-02 github

Driver 570.123.01 is actually a newer (after 570.124.04) release that does not include anything in 570.124.04 but specific Vulkan fixes. see https://github.com/NVIDIA/open-gpu-kernel-modules/releases/tag/570.123.01

Eekce 2025-03-02 github

I've tested with both. The 570.123.01 Vulkan drivers do indeed fix the issue but they are not in the Arch repos for whatever reason. The regular 570.124.04 drivers (i.e. the ones in the Arch repos) do not fix the issue.

Ttuxrinku 2025-03-03 github

I can't get the game to launch on 570.123.01. As soon as I start it I get a wine C++ assertion failed error. The file is loader_thunks.c. There is no problem with other games. I also tried clearing the vkd3d cache files but still nothing. When I go back to 570.124.04 + pyroveil there is no problem.
Any idea how to fix this ?

Mmrhanman 2025-03-05 github

I also tried 570.123.01 on Arch with the same results as tuxrinku. I tried a few different kernels and proton versions, but the error message was the same with every combination. I also tried "VKD3D_DISABLE_EXTENSIONS=VK_NV_low_latency2" as suggested for another game with the same error, but it did not help.

TTikilou 2025-03-05 github

Driver 570.123.01 is actually a newer (after 570.124.04) release that does not include anything in 570.124.04 but specific Vulkan fixes. see https://github.com/NVIDIA/open-gpu-kernel-modules/releases/tag/570.123.01

I've uploaded packages of the nvidia 570.123.01 beta driver for ArchLinux, Manjaro, CachyOS on a secondary sub-domain of my website, if some are interested, it should save you some time. 👍

I also provided the modified pkgbuild used to create the packages which were originally on AUR for nvidia-beta, nvidia-beta-dkms, and nvidia-utils-beta

http://redirect.pepinature.net/nvidia-ff7rebirth/

Original .run from nvidia is downloadable here : https://developer.nvidia.com/downloads/vulkan-beta-57012301-linux

PPolisanTheEasyNick 2025-03-05 github

I've uploaded packages of the nvidia 570.123.01 beta driver for ArchLinux, Manjaro, CachyOS on a secondary sub-domain of my website, if some are interested, it should save you some time. 👍

Thank you for your work!
Could you please provide lib32-nvidia-utils package too?
For me, I'm receiving removing nvidia-utils breaks dependency 'nvidia-utils=570.124.04' required by lib32-nvidia-utils error, on Arch Linux.

TTikilou 2025-03-05 github

I can't get the game to launch on 570.123.01. As soon as I start it I get a wine C++ assertion failed error. The file is loader_thunks.c. There is no problem with other games. I also tried clearing the vkd3d cache files but still nothing. When I go back to 570.124.04 + pyroveil there is no problem. Any idea how to fix this ?

Image

Same with the 570.123.01 driver, tested now. :-/

Eekce 2025-03-06 github

On Archlinux, you can install the Vulkan driver using this PKGBUILD.
https://github.com/Frogging-Family/nvidia-all
It will conflict with the egl-x11 package. So you'll need to install the packages by running pacman with --overwrite *egl* and uninstall the egl-x11 package to deal with that.

Hhivehivemind 2025-03-07 github

@hivehivemind We found another game sync bug. A workaround was pushed to vkd3d-proton master. HansKristian-Work/vkd3d-proton#2362 Hopefully that fixes the hang.

@HansKristian-Work Updating back, the latest Rebirth patch (Version 1.003) summed with this commit fixed the hanging in that area. Several minutes tested.

Thank you for such a quick action on it!

Ddarkpumpkins 2025-03-08 github

The game work with the PYROVEIL method but there is a lot of stutter into cut scenes and after the cut scene the game does stutter and the framerate collapse I don't know why ...

I have to go back to main menu and to load a save to fix the stutter + framerate issue everytime.

( The game work properly into 1080p under Windows )

My config :
CPU : Ryzen 7700
GPU : RTX 3060ti ( driver : 570.86.16 )
RAM : 32 Go DDR5 6000Mhz
OS : Ubuntu 24.10
Desktop : Wayland

SSloogs 2025-03-08 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2708399676

I had a similar issue. I fixed it by:

  • Disabling Variable Refresh Rate for all my monitors in my desktop settings
  • Switching the sync mode from VRR to V-Sync in game
  • Matching the monitor refresh rate to the locked FPS in game (59.95 Hz and 60 FPS for example)
Ddarkpumpkins 2025-03-08 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2708424698

I do not have VRR available on my screens, refresh rate is set at 60 Hz on my both screens and in the game the V-Sync was already on.

edit : following your hint around refresh rate I turned off my second screen and the game run way better ! ( still doesn't match Windows performance but playable at least in medium )

I do have 2 screens with not exactly the same refresh rate 60hz and 59.95hz and with different resolution 1920x1080 and 1920x1200 ( 16:10 ) for the other.

Kkorodarn 2025-03-08 github

I tried the solution of going to vulkan beta driver using frogging-family but I now get an Assertion failed! error from loader_thunks.c like [tuxrinku] earlier

Expression: "!status && vkQueuePresentKHR"

Not sure if there is a way to override this.

Ddabdude420 2025-03-08 github

still unplayable linux i have geforce rtx 4060
FINAL FANTASY VII REBIRTH - 2025-03-08 3-48-56 PM.mp4
https://drive.google.com/file/d/1rwe_jRqGmOjasmm-M9JmCu0GDQPdADxk/view?usp=drive_web

On Sat, Mar 8, 2025 at 2:28 PM korodarn @.***> wrote:

I tried the solution of going to vulkan beta driver using frogging-family
but I now get an Assertion failed! error from loader_thunks.c like
[tuxrinku] earlier

Expression: "!status && vkQueuePresentKHR"

Not sure if there is a way to override this.


Reply to this email directly, view it on GitHub
https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2708457988,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/BOZLYIWI7KAJTZ7COCGFF4D2TNAGDAVCNFSM6AAAAABVXRFUACVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOMBYGQ2TOOJYHA
.
You are receiving this because you commented.Message ID:
@.***>
[image: korodarn]korodarn left a comment (ValveSoftware/Proton#8408)
https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2708457988

I tried the solution of going to vulkan beta driver using frogging-family
but I now get an Assertion failed! error from loader_thunks.c like
[tuxrinku] earlier

Expression: "!status && vkQueuePresentKHR"

Not sure if there is a way to override this.


Reply to this email directly, view it on GitHub
https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2708457988,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/BOZLYIWI7KAJTZ7COCGFF4D2TNAGDAVCNFSM6AAAAABVXRFUACVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOMBYGQ2TOOJYHA
.
You are receiving this because you commented.Message ID:
@.***>

WWilsonMMM 2025-03-10 github

I originally came across this issue a couple of days ago when I was running a 3060Ti and the game had no game textures which I eventually got working with the pyroveil fix. I've since swapped to the radeon 9070 thinking it would be better...

I had the flashbang issue yesterday before realizing I needed to swap the Proton Experimental to bleeding edge for the issues to be fixed. I went to bed thinking the game was all sorted... and then I got to Kalm; I've never seen a sky so white.

Image

The game looks perfectly fine whenever the 'sun' isn't in play but the contrast is stark when it is. I've tried with and without HDR. In this case, the in-game brightness and HDR are set to 0 with my system brightness at 50~%.

OS: CachyOS (Fully up to date)
AMD Drivers: RADV(?) I've tried these and the AMDGPU ones
Proton: Experimental [bleeding-edge]
Launch options: gamescope -H 1440 -W 3440 -f -b --hdr-enabled -- %command%

Fairly new to Linux and especially the AMD GPU stack so this is a lot for me to get my head around but I thought I was in a good place. Can anyone confirm whether this is how the game is 'supposed' to look, whether I'm missing some config/setup, or whether there's something else that needs patching.

Edit:
Occurs to me that it might be useful to see my launch options

KKurinPawpad 2025-03-11 github

@WilsonMMM Does this occur outside of Kalm, such as on wide-open areas in the world map? Because I have noticed that my game is sometimes ridiculously bright in spots and ridiculously dark in others, if the camera angle is just wrong. It seems like it decides to base its exposure settings on random spots. There's a place in Under Junon that is pretty much 100% black all the time (behind the boats near the water).

Nnob8 2025-03-11 github

I'm observing some pretty consistent crashes in Cosmo Canyon, in the northern part of the map.

OS: OpenSUSE Tumbleweed
GPU+Driver: NVidia RTX 3080, Driver version 570.124.04-32.1
Proton: Hotfix, checked bleeding-edge too
Launch options: PROTON_LOG=1 PYROVEIL_CONFIG=/path/to/pyroveil/hacks/ffvii-rebirth-nvidia/pyroveil.json PYROVEIL=1 gamemoderun gamescope --hdr-enabled -h 1080 -r 120 -f -- %command%

I tried dropping gamescope and running in the Plasma DE (Wayland and X), but no luck. Turning on/off adaptive sync seems not to make a difference either. Monitor refresh rate is set to 120fps to match the game update rate.

The point in the log where the hang+crash happened was here (No lines skipped, the call to ID3D12Heap happens every couple milliseconds leading up to this):

766.706:0138:0234:warn:vkd3d-proton:d3d12_resource_QueryInterface: {6b3b2502-6e51-45b3-90ee-9884265e8df3} not implemented, returning E_NOINTERFACE.
886.879:0138:013c:warn:seh:OutputDebugStringW L"LowLevelFatalError [File:Unknown] [Line: 1246] \nGameThread timed out waiting for RenderThread after 120.00 secs\n"

Reproduction: Ride around near Spiritwastes tower and the chocobo stop to the west of it.

steam-2909400.log

WWilsonMMM 2025-03-11 github

@WilsonMMM Does this occur outside of Kalm, such as on wide-open areas in the world map? Because I have noticed that my game is sometimes ridiculously bright in spots and ridiculously dark in others, if the camera angle is just wrong. It seems like it decides to base its exposure settings on random spots. There's a place in Under Junon that is pretty much 100% black all the time (behind the boats near the water).

I'm still in Kalm. The graphics issues are extremely distracting so every time I try to play it I end up turning it off to search for fixes. Below are both from the entrance to the clocktower with the camera rotated slightly.

Image

Image

Image

Image

Mmbriar 2025-03-11 github

That doesn't look like a bug, that's just how the sun exposure effect looks in the game.

Eekce 2025-03-11 github

I'm pretty sure you're just describing the way the engine handles exposure. I'm pretty sure it's intentional (how the game is supposed to look) and you can see the effect done intentionally in some movies (e.g. windows being clipped while indoors and such). Final Fantasy VII Remake did this as well, here is a timestamped clip I just grabbed via youtube search https://youtu.be/NE0KGdUxzgg?t=489

PPolisanTheEasyNick 2025-03-11 github

I'm still in Kalm. The graphics issues are extremely distracting so every time I try to play it I end up turning it off to search for fixes.

Are we even sure that's Proton's issues and not Unreal Engine's?

WWilsonMMM 2025-03-11 github

I'm still in Kalm. The graphics issues are extremely distracting so every time I try to play it I end up turning it off to search for fixes.

Are we even sure that's Proton's issues and not Unreal Engine's?

Not at all, hahaha. It looks way over done but since I'd had a number of issues I honestly wasn't sure if it was another or simply how the game/engine was. @mbriar, @ekce, @PolisanTheEasyNick , Thanks for all confirming.

Cclement-analogue 2025-03-12 github

Hi,
I am affected by this bug. I have a GPU Nvidia 4090. After the opening scene, textures are missing, and I am unable to play. It looks like the screen shot.
This thread is huge, so I was not able to see if it exists a workaround.

Oopi3636 2025-03-12 github

Hi, I am affected by this bug. I have a GPU Nvidia 4090. After the opening scene, textures are missing, and I am unable to play. It looks like the screen shot. This thread is huge, so I was not able to see if it exists a workaround.

Nvidia is working a fix, but in the mean time you can use PYROVEIL to work around the bug.

Jjobs-git 2025-03-12 github

The fix was on 570.123.01 driver. Workaround for black screen in games is switching to Wayland but it will cause mouse lag. Switching to VRR to Vsync repeatedly helps alleviate the issue. However, with this driver, I am now getting assertion failed.

This did not work:

VKD3D_CONFIG=dxr VKD3D_FEATURE_LEVEL=12_2 VKD3D_SHADER_MODEL=6_6 %command%
Sshelterx 2025-03-15 github

Nvidia Vulkan Dev 570.123.06 was releaed yesterday. No idea if it gets rid of the assertion failed message but give it a try.
* Fix crash on vkQueuePresent under some circumstances on Linux

Ggianmarcotoso 2025-03-15 github

In my case, with Pyroveil on a 4090, everything works mostly fine but I'm currently experiencing a missing skybox in the Junon region (Chapter 4). The sky was correctly rendering in Chapter 2. I've tried several proton versions but nothing seems to change.

Image

Ggilvbp 2025-03-16 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2727040194

Did you try the last vulkan driver?

Ggianmarcotoso 2025-03-16 github

Replying to [#8408 (comment)](https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2727040194)

Did you try the last vulkan driver?

I just did, the issue persists. I've also tried manually deleting the VKD3D shader cache and verifying game files, to no avail.

Jjobs-git 2025-03-16 github

Replying to [#8408 (comment)](https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2727040194)

Did you try the last vulkan driver?

I just did, the issue persists. I've also tried manually deleting the VKD3D shader cache and verifying game files, to no avail.

This I learned the hard way, I created a new game and those corruptions where gone.

Using save file that I initially uses old driver seems to cause the graphics corruption to stick even though I have changed to new driver already.

Ggianmarcotoso 2025-03-16 github

I managed to copy the game and my saves to my Windows machine and I have the same issue, it's a bug of the game. I think my previous messages should be marked as off topic!

Hhobosock 2025-03-16 github

I've encountered an unusual problem using Pyroveil that I haven't seen anyone else report. I haven an Nvidia 3060 and had the same missing environment bug as everyone else. I installed the Pyroveil workaround as explained above and the game worked great, up until the most recent update. Now I have lots of missing textures, showing up as a uniform gray or black, missing skyboxes, and weird lighting effects. I tried reinstalling Pyroveil and have cleared the caches numerous times. Has anyone else experienced this?

EndeavourOS, Nvidia 3060, driver versions 570.124.01 and 570.124.04

Image
Image
Image
Image

WWilsonMMM 2025-03-17 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2713087034

I saw someone playing through Kalm on the PS5 and can confirm that the whiteout here is not intended. I'm now on the Junon side and have another incredible instance of it. In this case the sky is fine until you go up a wall into an event area.

Image

System wise, fully updated CachyOS, AMD GPU, using the Proton Experimental (Bleeding Edge) as of 2025-03-11.

Cclement-analogue 2025-03-17 github

I tried the nvidia driver 570.86.15, but it is not solving the issue. I guess the one in Ubuntu repo still does not have the bug fix.
However, PYROVEIL solves the issue, at least to start the game!

Bbinaryberserker 2025-03-18 github

PYROVEIL fixed FF7 Rebirth for me, at least so far! When first starting in Midgar, I was getting a transparent world except for the skybox, same as some others have shown.

I followed the build instructions for PYROVEIL, copied the FF7 Rebirth pyroveil.json to next to ff7rebirth.exe and added PYROVEIL=1 %command% to my launch options in steam.

I also had to switch my compatibility tool and force a specific one to get it to rebuild my shaders, i came back later on, after it rebuilt the shaders, and changed it back to the default and it still worked. so really it just needed to rebuild the shaders is all. There's probly a better way to do that, but it worked ¯\_(ツ)_/¯

Arch, RTX 3080, nvidia 570.124.04-1, linux 6.13.7.zen1-1, latest Proton Experimental

Sshelterx 2025-03-18 github

The Vulkan dev driver fixes the issue. It also fixes some other things so it's in fact better than the 570.124.06 driver.

ZZarkorix 2025-03-18 github

@shelterx
Sorry for my ignorance, but how do I install the dev driver? From where is it sourced?

Sshelterx 2025-03-18 github

@Zarkorix Depends on your distribution.
I have packages for CachyOS here. Might work on Arch too, haven't tested.

UPDATE:
Arch/CachyOS users can use this AUR now, it's being maintained again.
https://aur.archlinux.org/pkgbase/nvidia-vulkan

ZZarkorix 2025-03-18 github

@shelterx Ah - makes sense. Thank you. I'm on Fedora 41 at the moment. I assume at some point (relatively soon?) this will move out of beta and be released to everybody anyway.

Jjobs-git 2025-03-18 github

@shelterx Sorry for my ignorance, but how do I install the dev driver? From where is it sourced?

Warning: If you cant do step 1 then please dont install the dev nvidia for your system's safety

  1. Temporarily boot in run level 3. Just add 3 to the end of linux kernel command in grub.

  2. Remove existing nvidia driver

sudo apt autoremove *nvidia*
  1. Download the driver

Download this:

https://developer.nvidia.com/downloads/vulkan-beta-57012306-linux

  1. Install the driver
sudo ./NVIDIA-Linux-x86_64-570.123.06.run
  1. Reboot
AAlegrowin 2025-03-19 github

I'll wait for the dev build to be production grade, meanwhile:

Install and Configure Pyroveil for FF7 Rebirth

#!/bin/bash

git clone https://github.com/HansKristian-Work/pyroveil.git
cd pyroveil
git submodule update --init
cmake . -Bbuild -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$HOME/.local
ninja -C build install

config_url="https://raw.githubusercontent.com/HansKristian-Work/pyroveil/refs/heads/master/hacks/ffvii-rebirth-nvidia/pyroveil.json"
config_dest="$HOME/.steam/steam/steamapps/common/FINAL FANTASY VII REBIRTH/pyroveil.json"

curl -sLo "$config_dest" "$config_url"

if [ -f "$config_dest" ]; then
    echo "Pyroveil and configuration file installed successfully."
else
    echo "Failed to download configuration file."
fi

Don't forget to set steam launch argument, as pointed out before by @binaryberserker

PYROVEIL=1 %command%

Bonus Copy Saved data to get the bonus from Remake/Integrade

#!/bin/bash

src="$HOME/.steam/steam/steamapps/compatdata/1462040/pfx/drive_c/users/steamuser/Documents/My Games/FINAL FANTASY VII REMAKE"
dest="$HOME/.steam/steam/steamapps/compatdata/2909400/pfx/drive_c/users/steamuser/Documents/My Games/FINAL FANTASY VII REBIRTH"

if [ -d "$src" ]; then
  mkdir -p "$dest"
  cp -r "$src"/* "$dest"/
  echo "Save data copied successfully."
else
  echo "Source directory not found."
fi
ZZarkorix 2025-03-21 github

My distribution (Fedora 41) just updated to 570.133.07 and FFVII Rebirth now plays without issue on Proton Experimental - no missing textures, and no need for Pyroveil.

PPolisanTheEasyNick 2025-03-21 github

Arch Linux; 570.123.06 from nvidia-vulkan AUR package. RTX 3060. Can confirm that beta nvidia-vulkan drivers works very good without need of Pyroveil. Also I'm using optiscaler with updated DLSS .dll and with enabled frame generation via optiscaler. Game at 80% of time works flawlessly, but sometimes there are major slutters, which got fixed by some time if you go to the menu (like pause, where you can choose items for characters and so on, not main menu of game) and sit there while slutters gone. I have a feeling that that issue is performing because of no free video memory available (I can decrease chance of getting that issue if I close all gpu-memory affecting applications like discord and so on).

Image

Edit: updated driver to 570.133.07, all works as same. I can confirm that slutters was because of lack of GPU memory: while playing at medium texture setting, no slutters at all.
Noticed another issue: randomly game may freeze with end with need to kill process manually. Meanwhile at dmesg i can see message:

[ 6177.382226] NVRM: GPU at PCI:0000:26:00: GPU-165aaf64-98e5-4334-e0a2-f929f37136bf
[ 6177.382232] NVRM: Xid (PCI:0000:26:00): 31, pid=64338, name=ff7rebirth_.exe, Ch 000000d0, intr 00000000. MMU Fault: ENGINE GRAPHICS GPC1 GPCCLIENT_GCC faulted @ 0x26e0_00562000. Fault is of type FAULT_PDE ACCESS_TYPE_VIRT_READ

Tteknolog1k 2025-03-21 github

Still experiencing major stuttering and framerate drops on AMD when turning the camera, even w/ mesa-git. The radv_builtin_shaders directory does appear in my shadercache. Unsure of what to do, I'm still seeing drops of up to ~15fps when turning, even w/ VRR disabled and both my monitor and game capped at 60hz.

System Information:

  • CPU: Ryzen 7 7700X
  • GPU: Radeon RX 7700XT
  • Kernel: Linux 6.12.19
  • DE: KDE Plasma 6.3.3 (Wayland)
  • Distro: NixOS 25.05 "Warbler"
  • Using Proton Experimental [bleeding-edge]
Kkisak-valve maintainer 2025-03-26 github

Final Fantasy 7 Rebirth - Incorrect Game Version Being Installed

Issue transferred from https://github.com/ValveSoftware/Proton/issues/8563.
@Xipher-Zero posted on 2025-03-26T06:03:54:

System Information:

CPU: AMD 5700X3D
GPU: AMD 7900XTX
RAM: 32GB 3200mhz Memory
Storage: NVMe M.2 Drive

System Software Information:

Distro: Bazzite Stable (Based on FC41)
Kernel: 6.13.7-108.bazzite.fc41.x86_64
Graphics Driver: Mesa 25.0.2
Steam Version: 1741737356
Compatibility Layer Used: Proton9, GE-Proton9-26

Problem Description:

I purchased FF7 Rebirth on launch and have been playing the game at 3200x1800@90hz without issue with all settings on high. The game does not have any problems running on my hardware.

Square Enix provided an update (v.1.0.0.3) for FF7 Rebirth to accommodate the Steam Deck which included several fixes as well as platform optimizations for the hand held, but also REMOVED all hiDPI texture packages to reduce the games install footprint.

The problem is that systems like mine which have highly capable hardware are ALSO receiving ONLY the cut down version of the game with all hiDPI textures removed.

At this time there does not appear to be any method to get the steam client to provide the normal version of the game. (No beta channels available to override the cutdown version with the normal version)

This has greatly degraded the experience on my hardware outputting to a 65" OLED display as the game now looks HORRIBLE compared to how it ran just a few weeks ago. Frankly this is not an acceptable situation.

Expectation:

Provide a solution which allows Linux users with capable hardware to download / install the FULL game and not the striped down version which was created for the Steam Deck and other handheld hardware.

DDreynbolt 2025-03-27 github

Bazzite 41, game is rendering properly without Pyroveil now, but a new issue is happening: random freezes.

Driver 570.133.07 on a 4070 Super, on kernel 6.13.7-108.bazzite.fc41.x86_64

dmesg:

[ 1145.596300] NVRM: Xid (PCI:0000:01:00): 31, pid=10512, name=ff7rebirth_.exe, Ch 0000007c, intr 00000000. MMU Fault: ENGINE GRAPHICS GPC1 GPCCLIENT_GCC faulted @ 0xffffbbc0_0073b000. Fault is of type FAULT_PDE ACCESS_TYPE_VIRT_READ
XxXJSONDeruloXx 2025-04-02 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2754226829

Wanted to back this up and express the same frustration. I'd hope something like a per game toggle for "steam deck optimized version" opt out selection. I also have a 7900xtx and was wondering why the game looks so rough now. Thanks for pointing this out

BBillli11 2025-04-02 github

@xXJSONDeruloXx Are you also using Bazzite?
FF7Reirth is still 153.37GB on my machine.

XxXJSONDeruloXx 2025-04-02 github

@xXJSONDeruloXx Are you also using Bazzite? FF7Reirth is still 153.37GB on my machine.

Yes, Bazzite 41 but it is a -deck build, as I like the controller driven UI, is yours standard desktop?

BBillli11 2025-04-02 github

@xXJSONDeruloXx Yes. (arch linux)

Seem like steam treat Bazzite as a steam deck and downloaded the steam deck version.
Or just because I downloaded the game before the update.

XxXJSONDeruloXx 2025-04-02 github

any idea what steam uses to determine this? maybe I could edit a line in a config file somewhere to work around

XxXJSONDeruloXx 2025-04-02 github

It seems bazzite deck images pass in launch arg -steamdeck with steam to accommodate BPM game mode flags.

I was able to directly launch Steam with /usr/bin/steam to bypass and got the current slightly different client, which then let me start the full sized download of 144 gb

Still, this is not ideal in practice and should be clearly available to opt out

BBillli11 2025-04-02 github

@xXJSONDeruloXx Well this is opt-in (the -steamdeck arg, it's just opted-in by bazzite, not you).
May be you should open a issue to bazzite about setting steamdeck mode in the steam client?

SShin-Aska 2025-04-02 github

I am using Ubuntu 22.04 and I get the full sized client (153gb according to my steam client). I believe this is a bazzite issue more than anything.

Hhivehivemind 2025-04-02 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2771343073

This is offtopic to Proton.

In any case, this happens due to the "-steamdeck" launch argument, as some people stated, and it requires sudo to edit out of the gamescope session, which can be complicated with Bazzite's immutable image.

On CachyOS, which runs gamescope-session-plus, you can edit using sudo /usr/share/gamescope-session-plus/sessions.d/steam and around line 100 remove -steamdeck in the CLIENTCMD environment variable.
Everything normally works from there as far as I tested, still, and without the steam deck configuration bias for the games.

Hhivehivemind 2025-04-02 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2772342165

Testing further, mangohud is not available without the "-steamdeck" argument in game mode as well as frame limit, VRR, and every other game-specific feature (including FSR/NIS image scaling), so it is not something Bazzite would like to take out but maybe workaround that eventually. VRR will be set automatically depending on your display, but HDR can be changed in the Steam "Display" menu.
I would recommend either sticking with the "-steamdeck" argument or taking it out only if these changes don't bother you.

Edit: the game is very stuttery here using the full texture package in a RX 7900 XT and 5800X3D. This is disappointing and wasn't an issue on previous versions using the same game full size.

XXipher-Zero 2025-04-02 github

Wow progress! I went ahead and was able to update to the full version of FFVII Rebirth by running steam without it using any arguments from desktop mode (/usr/bin/steam) as was mentioned above. I will be locking this version using some of the techniques to stop steam from updating a single game in the future.

I'm actually very surprised that we aren't hearing of people using something like ChimeraOS or Nobara reporting the same problem as I'm fairly certain they also use these flags to provide a console experience on general purpose hardware. It could also just be that there are more of us using Bazzite than either of these other systems.

I still feel that there should be an option within steam if there are 2 branches of a game released (Normal and Steamdeck) that allows the user to manually select a branch if their configuration is outside of Valves or the developers expectations.

Jjorbuedo 2025-04-02 github

The amount of people using linux is low, using a distro that automatically adds that flag on a desktop would be even lower (I use endevour and never had that), from those only a few would notice the texture size being wrong, and then being vocal enough to say something online. So yeah, it's to be expected not a lot of reports too fast.

As for the nvidia driver, nice to see 570.133.07 is installed already on Arch. Anyone checked the performance vs the pyroveil hack?

PPolisanTheEasyNick 2025-04-02 github

Anyone checked the performance vs the pyroveil hack?

Can't tell about performance in FPS, but when using Pyroveil + Optiscale's frame generation, I experienced full freezes. However, without Pyroveil and on 570.133.07 driver, there were no freezes with frame generation enabled.
Note that there are still freezes due to invalid memory accesses, especially in the Cosmo Canyon region for me, but I believe they are not related to frame generation but related to bugs in game on engine (why they occure muuuch much likely in specified region then?).
At this state I'm happy how playable this game is for me: >100 FPS in 3060 with framegen; sometimes full freezes, but it requires closing all GPU-related apps (Discord, Spotify, even Firefox) for freeing as much GPU memory as possible. If there is no much GPU memory available, then massive slutters occurs :c

XxXJSONDeruloXx 2025-04-02 github

I noted that if I then switched to game mode (where -steamdeck arg is passed in gamescope session config) it would automatically start trying to DL the 20 gigs from the SteamDeck Depot and overwrite the full version of the game.

Jjobs-git 2025-04-19 github

After Trial in Chapter 13, Cloud movement bugs out, like however to walk the game stucks you in that position permanently.

Using Steam Flatpak Proton Experimental Nvidia RTX 4070 570.123.06 Ubuntu 24.04

Jjobs-git 2025-05-06 github

GUI freezing when capturin Red XIII photo using quest-camera during Writer's block odd-job in Gongga beach.

Ubuntu 24.04 + RTX 4070 S + 575.51.03 + Proton Experimental

Here is the SS when it happens, notice the GUI for the quest-camera did not load:

Image

AABrony 2025-05-12 github

I was having terrible frametimes and fps drop on specific camera placements on AMD RX 9070XT with latest proton and mesa. It wasn't even running without specifying some flags mentioned above.

I'm glad to tell - issue is completely fixed by replacing amdvlk package with vulkan-radeon (in case of conflict install vulkan-radeon first, then you'll be able to remove amdvlk). It also improved my perfomance in different games by A LOT. Maybe this would help you too. Arch wiki does mention this, but i somehow overlooked it.

Ddabdude420 2025-06-02 github

ff7 rebirth work now

On Sun, Jun 1, 2025 at 6:04 PM André @.***> wrote:

andreldmonteiro left a comment (ValveSoftware/Proton#8408)
https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2927975748

I'm having problems with audio. When playing the story so far video from
main menu using a 5.1 surround sound system, the audio supposed to be
coming from the centre speaker is played instead on the rear left speaker.


Reply to this email directly, view it on GitHub
https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-2927975748,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/BOZLYITDONK52LL7DIRRXVL3BN2FLAVCNFSM6AAAAABVXRFUACVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDSMRXHE3TKNZUHA
.
You are receiving this because you commented.Message ID:
@.***>

SShochraos 2025-06-27 github

I am getting absurdly long loading times on CachyOS with Proton Experimental as well as Proton 9.0.4.

During these loading times my CPU usage ramps up to 100% and then down to normal usage 3 times per load. The loading screens usually last 3-5 minutes.

Ryzen 7 7800x3D
NVIDIA Geforce RTX 4090
32GB 6000 CL30
Lexar NM790 SSD (I am running on an SSD)

The proton log repeatedly prints the following sequence for around 1000 lines until the loading finishes:

10047.985:015c:0260:warn:vkd3d-proton:d3d12_command_queue_QueryInterface: {db6f6ddb-ac77-4e88-8253-819df9bbf140} not implemented, returning E_NOINTERFACE.

10047.989:015c:0278:warn:vkd3d-proton:vkd3d_native_sync_handle_release: Failed to release semaphore (#12a).

10047.989:015c:0278:warn:vkd3d-proton:dxgi_vk_swap_chain_wait_worker: Failed to increment swapchain semaphore. Did application forget to acquire?

Anybody having the same issues as me?

DDids 2025-07-01 github

Will this ever fully work on non-Steam Deck hardware while running something like bazzite-nvidia-deck with SteamDeck=0? Or is it entirely up to the developer to just not force low resolution textures and other Steam Deck specific limitations?

Pperru 2025-07-02 github

Lot of Out of memory at Costa del Sol, can't do some side quests because of this...

Costa del Sol make the game crash so often and some side quests AND main quest after chapter 12 can't be continue because of this...

Anything else is perfect.

OS:ArchLinux
GPU:AMD Radeon RX 6900 XT
CPU:AMD Ryzen 9 7950X3D
RAM:32 GB @7200Mhz

Bbluca 2025-07-19 github

I am seeing sporadic texture issues with a Radeon card. This seems different from the issue reported earlier on Nvidia.

The game works fine for most of the time, but after some time that varies between 1 and 2 hours textures start to become corrupt, first faintly and then more and more. This usually first affects the terrain, as in the screenshot, but then also affects other objects like buildings and vehicles after a while. Reloading the latest save fixes the issue.

Example:

Image

Radeon 9060 XT
Debian Testing
kernel 6.15.6
amdgpu firmware 20250708
mesa 25.1.5
proton experimental

Kkisak-valve maintainer 2026-02-04 github

Final Fantasy VII Rebirth

Issue transferred from https://github.com/ValveSoftware/Proton/issues/9465.
@Capricornucopia posted on 2026-02-04T01:41:12:

Compatibility Report

  • Name of the game with compatibility issues: Final Fantasy VII Rebirth
  • Steam AppID of the game: 2909400

System Information

  • GPU: NVIDIA GeForce RTX 3050 Ti Laptop GPU
  • Video driver version: xserver-xorg-video-nouveau
  • Kernel version: 6.17.0.8-generic
  • Link to full system information report as Gist:
  • Proton version: 9.0-4 and experimental

I confirm:

  • [x] that I haven't found an existing compatibility report for this game. X
  • [x] that I have checked whether there are updates for my system available. X
Proton: 1749140930 proton-9.0-4f
SteamGameId: 2909400
Command: ['/home/crowgirl/snap/steam/common/.local/share/Steam/steamapps/common/FINAL FANTASY VII REBIRTH/ff7rebirth.exe']
Options: {'forcelgadd'}
depot: 3.0.20251216.191774
pressure-vessel: 0.20251210.0 scout
scripts: 0.20251210.0
sniper: 3.0.20251216.191774 sniper 3.0.20251216.191774
Kernel: Linux 6.17.0-8-generic #8-Ubuntu SMP PREEMPT_DYNAMIC Fri Nov 14 21:44:46 UTC 2025 x86_64
Language: LC_ALL None, LC_MESSAGES None, LC_CTYPE None
Effective WINEDEBUG: +timestamp,+pid,+tid,+seh,+unwind,+threadname,+debugstr,+loaddll,+mscoree
======================

Symptoms

Game doesn't even launch.

steam-2909400.log

Reproduction

Kkisak-valve maintainer 2026-02-04 github

Hello @Capricornucopia, a header only Proton log hints that something went awry before the game itself had a chance to start. If you close Steam, then run steam from a terminal, does the terminal spew give a hint when you tell Steam to start the game? Alternatively, you can watch console_log.txt in ~/.local/share/Steam/logs (or the equivalent for the snap packaging).

CCapricornucopia 2026-02-04 github

This is the output of "steam" at the BASH command line:

crowgirl@crowgirlscrowsnest:~$ steam
Testing for explicit PulseAudio choice...
...and PulseAudio has been explicitly chosen, so using it.
INFO: filtering /home/crowgirl/.local/share/flatpak/exports/share out of XDG_DATA_DIRS because it is unreachable
INFO: filtering /var/lib/flatpak/exports/share out of XDG_DATA_DIRS because it is unreachable
INFO: filtering /usr/share/kubuntu-default-settings/kf5-settings out of XDG_CONFIG_DIRS because it is unreachable

And the Steam client GUI (flatpack) does launch.

I cannot find a console_log.txt file, nor any logs for Steam in general except for the one steam-2909400.log file that was made for Final Fantasy VII Rebirth because of the special launching instructions given to Proton?

Thank you so much for your help. I'm not great at coding, but I do have a good general sense of Linux and computing in general as a cybersecurity professor. If you use any terminology that I don't understand, I will ask.

I feel fairly comfortable at the BASH command line, but I usually have to look up commands that I use less often.

Ssimifor 2026-02-18 github

@Capricornucopia ~/.local/share/Steam/logs is used by steam when installed from your package manager, flatpak doesn't usually let programs have direct access to your home folder, so the steam folder would be placed somewhere under ~/.var

Aarnxxau 2026-02-22 github

This game still suffers from abnormal stuttering that isn't present when playing on the same hardware on Windows (at least on AMD).

It fails to maintain a stable framerate even when slowly panning the camera

Is anyone else experiencing this?

Jjbbanegas 2026-02-23 github

This game still suffers from abnormal stuttering that isn't present when playing on the same hardware on Windows (at least on AMD).

It fails to maintain a stable framerate even when slowly panning the camera

Is anyone else experiencing this?

Yes, I'm also on AMD (9070XT) and the game stutters on certain points no matter the framerate cap. I tried it on Windows recently and the stutter was not present at the same moments, but there was screen tearing when it stutters on Linux, even with VRR active.
Something isn't right with this game on PC.

AArthurHenriqueVieira 2026-05-20 github

This game still suffers from abnormal stuttering that isn't present when playing on the same hardware on Windows (at least on AMD).

It fails to maintain a stable framerate even when slowly panning the camera

Is anyone else experiencing this?

What distro are you using? if you're on any that simulates the SteamOS, like bazzite for HTPC, CachyOS Handheld, ChimeraOS, the game have a smaller download size due to thinking it's a Steam Deck, and I notived that version stutters like crazy when moving the camera, but on the full size game it doesn't

On this note, anyone found a fix for this steamdeck smaller size?

RRTbecard 2026-05-20 github

This game still suffers from abnormal stuttering that isn't present when playing on the same hardware on Windows (at least on AMD).

It fails to maintain a stable framerate even when slowly panning the camera

Is anyone else experiencing this?

Yup. Pop os 24.04 (gnome DE); rx 9070xt; ryzen 5 9600x; using optiscaler for FSR 4.0.2.

I had really bad camera pan stutter when my background model detail was set to ultra. Lowering that to high greatly mitigated the issue, and made it very playable. Stutter is still present, but quite rare now.

My CPU and GPU were nowhere near bottlenecking when i had the stutter issues... So i assumed it was poor game optimization.

Jjobs-git 2026-05-20 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-4502365415

does this also occur notoriously in Nibelheim with Cloud + Seph?

AArthurHenriqueVieira 2026-05-21 github

Replying to [#8408 (comment)](https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-4502365415)

does this also occur notoriously in Nibelheim with Cloud + Seph?

Nibelheim is one of the worse performance places in the game in my experience

Kkisak-valve maintainer 2026-06-24 github

Proton Experimental does not enable FSR4 on RDNA4 correctly.

Issue transferred from https://github.com/ValveSoftware/Proton/issues/9908.
@Hupro posted on 2026-06-24T23:19:43:

steam-2909400.log

Using the latest Proton Experimental, when FSR4 is enabled in game on RDNA4, it enables FSR4.1.1 however it using the FSR3 fallback. This was confirmed using optiscaler to use the upscaler watermark. By default the game loads FSR4.1.1 however the watermark shows it is actually using FSR3. Manually switching the game to FSR4.1.0 (using optiscaler) shows the correct MLSR watermark.

I've attached 2 images to illustrate this issue. Unsure if it is helpful in this case but I've also attached the proton log when booting a game and the incorrect upscale model is chosen.

Image Image
Zziobroaatwit 2026-07-10 github

This game still suffers from abnormal stuttering that isn't present when playing on the same hardware on Windows (at least on AMD).

It fails to maintain a stable framerate even when slowly panning the camera

Is anyone else experiencing this?

Yes, same issue here, 5800X3D, 7900XTX on SteamOS 3.8.22, and Proton Experimental, this game stutters so unbearably bad on camera rotations that playing it on PC is worthless. Capping the FPS does not "Fix" the problem because it has noticable stuttering even when a 30FPS cap is engaged in Game Mode QAM that presents as ~28 or so FPS when it happens.

Turning the camera:
Image

Stationary:
Image

A loss of almost 20FPS and egregious stuttering seems completely out of the ordinary and should probably be reviewed. On top of the stuttering assets in various locations can be seen popping in when they do not even on PS5 with a paltry Ryzen 3600 equiv CPU. The game is installed on a Gen 4 NVME.

WwillWallace-RIT 2026-07-18 github

I finally got it running! It appears a freshie start doesn't work in proton 9+. it might not have been caught due to getting past the initial menu into the starting cut scene but it freezes up just before game play. I have haven't had a chance to check the diagnostic yet,but I am curious if this is a whack-a-mole (as opposed to whack-a-box) situation where switching between proton versions gets one past certain parts but not subsequent others. Maybe this is call for a dynamic version switching mode for the compatibility layer based on aggregate UE?(???) EDIT: truth be told, Gemini said 8; hotfix, exp, 11,& 10 didn't work for me at this moment.

WwillWallace-RIT 2026-07-22 github

I sent this one to square because it is a little more in-depth but has anyone else notice thread timing issues where it works but is wonky in the change up from battle to exploration. I didn't notice it on the PS5 and Linux might be handled threads and locks a little differently than Windows or BSD,,,

Aarzardk 2026-08-08 github

EndeavourOS, Linux 7.1.5-arch1-2, mesa 26.1.6, Radeon RX7900XTX, Ryzen 7 7800X3D, Proton 11.0

White indicator appears mostly on wet surfaces but also appearing on some other surfaces (Nibelheim). Is it bug or feature?

Things i tried:
Changing Proton version to CachyOS_Latest
Disabling Russian localization (mod)
Changing graphical preset to "Low"
Changing upscaling method from FSR to TAAU

Video

TTxordi00 2026-08-11 github

Steam Deck, main Steam OS branch, latest Proton Experimental.

Game launches properly and on new game it crashes after the first video. If you watch the video, it crashes when the graphics engine engages. If you skip the video, then it crashes during the loading screen. I do not have access to any logs right now, but I can upload them later on. Anyone else has encountered this issue? It happens with the latest Proton-CachyOS and Proton-GE as well.

TTxordi00 2026-08-11 github

It happens only from gamescope. From the Plasma session the gameplay starts without issues. Here's the log from gamescope.
steam-2909400.log

Ssimifor 2026-08-11 github

@Txordi00 the game hasn't crashed on my end, can you share your display/graphics settings?

TTxordi00 2026-08-11 github

@Txordi00 the game hasn't crashed on my end, can you share your display/graphics settings?

I kept everything on default.

IMG_20260812_012904_535.jpg

IMG_20260812_012859_550.jpg

TTxordi00 2026-08-11 github

@Txordi00 the game hasn't crashed on my end, can you share your display/graphics settings?

Have you tried a new game & from gamescope?

Ssimifor 2026-08-11 github

@Txordi00 all my attempts were new games started from steamos game mode, the initial cinematic plays fine and after it ends I can control zack fair, I don't play much past this point as it sounds like this is where it crashes for you. Your logs show that the gpu is crashing. Have you rebooted your deck after first running into this? switching between game and desktop mode doesn't reboot the device, for example.

TTxordi00 2026-08-12 github

@Txordi00 all my attempts were new games started from steamos game mode, the initial cinematic plays fine and after it ends I can control zack fair, I don't play much past this point as it sounds like this is where it crashes for you. Your logs show that the gpu is crashing. Have you rebooted your deck after first running into this? switching between game and desktop mode doesn't reboot the device, for example.

Yes, I have tried to reboot but to no avail.

Ssimifor 2026-08-12 github

@Txordi00 what os build are you on? I updated to main specifically to check this, so there's a chance I'm on a newer build than you, and it happens to already be fixed. I'm on 20260807.1000

Also, what is your deck model? It shouldn't matter most of the time, but still would be good to specify in this case. I'm on deck oled.

Update:
A few other things. After you get another crash, go to steam > settings > system> create report (towards the end of the list) > save to desktop
This will create a text file on your desktop, upload it here as well

TTxordi00 2026-08-13 github

Replying to https://github.com/ValveSoftware/Proton/issues/8408#issuecomment-5269483559

Hi Simon! Sorry for taking long to answer. I do not have access now to the steam deck. I'm on the oled and on the same build than yours. I'll do that and upload the file tonight. I posted the issue in the gamescope repository as well. Thanks for your help!

TTxordi00 2026-08-13 github

Now it goes through the first part and it crashes later on... I'm starting to think that there's something wrong with my Steam Deck.
systemreport-20260813212313.txt

Proton versions

Launch options

Launch lines

Upstream links

DLLs