Smoke not rendering & visual glitches.

[ecabott]

Your system information

• Steam client version (build number or date): latest from default multilib repo
• Distribution (e.g. Ubuntu): Archlinux
• Opted into Steam client beta?: [Yes/No] No
• Have you checked for system updates?: [Yes/No] Yes
• Steam Logs: [generate by running this command in a terminal tar -zcvf ~/Desktop/steam-logs.tar.gz ~/.steam/steam/logs]
• GPU: Nvidia/Intel/AMD Nvidia

Please describe your issue in as much detail as possible:

Describe what you expected should happen and what did happen. Please link any large code pastes as a Github Gist

Steps for reproducing this issue:

1. Smoke
2. Smoke
3. Smoke
   
   
   
   
   
   
   

[danginsburg]

Can you please post Steam -> Help -> System Information from the machine? We need more info on the GPU and driver to be able to reproduce it. Thanks.

[ecabott]

Computer Information:
Manufacturer: LENOVO
Model: LNVNB161216
Form Factor: Laptop
No Touch Input Detected
Processor Information:
CPU Vendor: GenuineIntel
CPU Brand: Intel(R) Core(TM) i5-9300H CPU @ 2.40GHz
CPU Family: 0x6
CPU Model: 0x9e
CPU Stepping: 0xa
CPU Type: 0x0
Speed: 4100 MHz
8 logical processors
4 physical processors
Hyper-threading: Supported
FCMOV: Supported
SSE2: Supported
SSE3: Supported
SSSE3: Supported
SSE4a: Unsupported
SSE41: Supported
SSE42: Supported
AES: Supported
AVX: Supported
AVX2: Supported
AVX512F: Unsupported
AVX512PF: Unsupported
AVX512ER: Unsupported
AVX512CD: Unsupported
AVX512VNNI: Unsupported
SHA: Unsupported
CMPXCHG16B: Supported
LAHF/SAHF: Supported
PrefetchW: Unsupported
Operating System Version:
"Arch Linux" (64 bit)
Kernel Name: Linux
Kernel Version: 6.11.5-arch1-1
X Server Vendor: The X.Org Foundation
X Server Release: 12401004
X Window Manager: GNOME Shell
Steam Runtime Version: steam-runtime_0.20240806.97925
Video Card:
Driver: Intel Mesa Intel(R) UHD Graphics 630 (CFL GT2)
Driver Version: 4.6 (Compatibility Profile) Mesa 24.2.5-arch1.1
OpenGL Version: 4.6
Desktop Color Depth: 24 bits per pixel
Monitor Refresh Rate: 59 Hz
VendorID: 0x10de
DeviceID: 0x1f91
Revision Not Detected
Number of Monitors: 1
Number of Logical Video Cards: 2
Primary Display Resolution: 1920 x 1080
Desktop Resolution: 1920 x 1080
Primary Display Size: 13.39" x 7.48" (15.31" diag), 34.0cm x 19.0cm (38.9cm diag)
Primary VRAM Not Detected
Sound card:
Audio device: Realtek ALC257
Memory:
RAM: 15864 Mb
VR Hardware:
VR Headset: None detected
Miscellaneous:
UI Language: English
LANG: en_US.UTF-8
Total Hard Disk Space Available: 487360 MB
Largest Free Hard Disk Block: 315292 MB
Storage:
Number of SSDs: 1
SSD sizes: 512G
Number of HDDs: 0
Number of removable drives: 0

[huupoke]

I also experience significant FPS drop and invisible smokes on Thera map.
I tried to reproduce this in loopback server (with map de_thera) but I can't. I currently only experience this when playing on official server. I also see my VRAM usage is near 100% (my GPU has 4GB VRAM).

System information

Computer Information:
Manufacturer: LENOVO
Model: LNVNB161216
Form Factor: Laptop
No Touch Input Detected
Processor Information:
CPU Vendor: AuthenticAMD
CPU Brand: AMD Ryzen 5 5600H with Radeon Graphics
CPU Family: 0x19
CPU Model: 0x50
CPU Stepping: 0x0
CPU Type: 0x0
Speed: 4279 MHz
12 logical processors
6 physical processors
Hyper-threading: Supported
FCMOV: Supported
SSE2: Supported
SSE3: Supported
SSSE3: Supported
SSE4a: Supported
SSE41: Supported
SSE42: Supported
AES: Supported
AVX: Supported
AVX2: Supported
AVX512F: Unsupported
AVX512PF: Unsupported
AVX512ER: Unsupported
AVX512CD: Unsupported
AVX512VNNI: Unsupported
SHA: Supported
CMPXCHG16B: Supported
LAHF/SAHF: Supported
PrefetchW: Unsupported
Operating System Version:
Debian GNU/Linux 12 (bookworm) (64 bit)
Kernel Name: Linux
Kernel Version: 6.1.0-26-amd64
X Server Vendor: The X.Org Foundation
X Server Release: 12101007
X Window Manager: GNOME Shell
Steam Runtime Version: steam-runtime_0.20240806.97925
Video Card:
Driver: NVIDIA Corporation NVIDIA GeForce RTX 3050 Ti Laptop GPU/PCIe/SSE2
Driver Version: 4.6.0 NVIDIA 535.183.01
OpenGL Version: 4.6
Desktop Color Depth: 24 bits per pixel
Monitor Refresh Rate: 165 Hz
VendorID: 0x10de
DeviceID: 0x25e0
Revision Not Detected
Number of Monitors: 1
Number of Logical Video Cards: 1
Primary Display Resolution: 1920 x 1080
Desktop Resolution: 1920 x 1080
Primary Display Size: 13.54" x 7.64" (15.51" diag), 34.4cm x 19.4cm (39.4cm diag)
Primary Bus: PCI Express 16x
Primary VRAM: 4096 MB
Supported MSAA Modes: 2x 4x 8x 16x
Sound card:
Audio device: Nvidia GPU a0 HDMI/DP
Memory:
RAM: 15842 Mb
VR Hardware:
VR Headset: None detected
Miscellaneous:
UI Language: English
LANG: en_GB.UTF-8
Total Hard Disk Space Available: 402059 MB
Largest Free Hard Disk Block: 139507 MB
Storage:
Number of SSDs: 1
SSD sizes: 512G
Number of HDDs: 0
Number of removable drives: 0

[W33X]

Same here, I'm on Nvidia

[tiagotriques]

Had the same problem twice in a game.

The smoke shadows were rendered, but the smoke "cloud" wasn't, depending on the camera angle.

The bug does not occur when watching the demo, only happened in game.

System info

Computer Information:
Manufacturer: Micro-Star International Co., Ltd.
Model: B560M-A PRO (MS-7D20)
Form Factor: Desktop
No Touch Input Detected
Processor Information:
CPU Vendor: GenuineIntel
CPU Brand: 11th Gen Intel(R) Core(TM) i5-11400F @ 2.60GHz
CPU Family: 0x6
CPU Model: 0xa7
CPU Stepping: 0x1
CPU Type: 0x0
Speed: 4400 MHz
12 logical processors
6 physical processors
Hyper-threading: Supported
FCMOV: Supported
SSE2: Supported
SSE3: Supported
SSSE3: Supported
SSE4a: Unsupported
SSE41: Supported
SSE42: Supported
AES: Supported
AVX: Supported
AVX2: Supported
AVX512F: Supported
AVX512PF: Unsupported
AVX512ER: Unsupported
AVX512CD: Supported
AVX512VNNI: Supported
SHA: Supported
CMPXCHG16B: Supported
LAHF/SAHF: Supported
PrefetchW: Unsupported
BMI1: Supported
BMI2: Supported
F16C: Supported
FMA: Supported
Operating System Version:
"Fedora Linux 41 (KDE Plasma)" (64 bit)
Kernel Name: Linux
Kernel Version: 6.12.7-200.fc41.x86_64
X Server Vendor: The X.Org Foundation
X Server Release: 12401004
X Window Manager: KWin
Steam Runtime Version: steam-runtime_0.20241024.105847
Client Information:
Version: 1733265492
Browser GPU Acceleration Status: Disabled
Browser Canvas: Unavailable
Browser Canvas out-of-process rasterization: Disabled
Browser Direct Rendering Display Compositor: Disabled
Browser Compositing: Disabled
Browser Multiple Raster Threads: Enabled
Browser OpenGL: Disabled
Browser Rasterization: Disabled
Browser Raw Draw: Disabled
Browser Skia Graphite: Disabled
Browser Video Decode: Disabled
Browser Video Encode: Disabled
Browser Vulkan: Disabled
Browser WebGL: Unavailable
Browser WebGL2: Unavailable
Browser WebGPU: Disabled
Browser WebNN: Disabled
Video Card:
Driver: NVIDIA Corporation NVIDIA GeForce RTX 2060/PCIe/SSE2
Driver Version: 4.6.0 NVIDIA 565.77
Desktop Color Depth: 24 bits per pixel
Monitor Refresh Rate: 59 Hz
VendorID: 0x10de
DeviceID: 0x1e89
Revision Not Detected
Number of Monitors: 1
Number of Logical Video Cards: 1
Primary Display Resolution: 1440 x 900
Desktop Resolution: 1440 x 900
Primary Display Size: 16.06" x 10.04" (18.94" diag), 40.8cm x 25.5cm (48.1cm diag)
Primary VRAM: 6144 MB
Sound card:
Audio device: USB Mixer
Memory:
RAM: 15861 Mb
VR Hardware:
VR Headset: None detected
Miscellaneous:
UI Language: English
LANG: pt_BR.UTF-8
Total Hard Disk Space Available: 227311 MB
Largest Free Hard Disk Block: 131155 MB
Storage:
Number of SSDs: 7
SSD sizes: 500G,240G,0B,0B,0B,0B,0B
Number of HDDs: 0
Number of removable drives: 0

[tiagotriques]

Happened once again, now I have it on video: https://np.hjkl.page/w.mp4

[danginsburg]

Hello,

We have been struggling mightily to reproduce this locally on any of our Linux boxes. If anyone is willing to help us out, we could really use a RenderDoc capture of the bug when it happens. Here are some instructions on how to capture cs2 with RenderDoc - if you can do this when the smoke disappears and send us the trace we're hopeful we can diagnose the root cause.

1. Download RenderDoc v1.36 for Linux from https://renderdoc.org/
2. Untar and run renderdoc_1.36/bin/qrenderdoc
3. Click the Launch Application tab and click "Warning: Vulkan capture is not configured. Click here to set up Vulkan capture."
4. Click yes to register the layer and enter your root password at the terminal. Now you can exit qrenderdoc
5. IMPORTANT: Download librenderdoc.tar.gz from https://www.dropbox.com/scl/fi/a9v9p6babramolb0yq911/librenderdoc.so.tar.gz?rlkey=d5umdtljlxixg55n9hh1kchxn&st=gfb0ouho&dl=0. This is a custom copy that works around a crash bug capturing cs2 with RenderDoc v1.36 (we are still working on landing a proper fix).

• tar xvf librenderdoc.tar.gz
• Overwrite the copy of librenderdoc.so in renderdoc_1.36/lib/

6. From a command line, launch cs2 as follows:
   ~/.steam/steam/steamapps/common/Counter-Strike Global Offensive/game$ ENABLE_VULKAN_RENDERDOC_CAPTURE=1 ~/.steam/steam/steamapps/common/SteamLinuxRuntime_sniper/run -- ./cs2.sh
7. You should see the RenderDoc overlay ("Capturing Vulkan" in the upper left corner of the screen)
8. When you see the bug to capture it, press F12. There will be a long pause and RenderDoc will write the capture to /tmp/RenderDoc
9. Send us the .rdc file written to /tmp/RenderDoc

In addition to renderdoc, if you can add -con_logfile to your command line and send us the contents of ~/.steam/steam/steamapps/common/Counter-Strike Global Offensive/game/csgo/console.log that would be helpful.

[tiagotriques]

Its not always that the bug happens on my machine. I tried to reproduce it but no luck so far :(

[danginsburg]

Its not always that the bug happens on my machine. I tried to reproduce it but no luck so far :(

Yes, we've had no luck reproducing it here either. One other detail: would you please post a screenshot of your Video Settings? That might help us repro. Thanks.

[huupoke]

@danginsburg
I have taken 3 samples on 3 separate smokes on a Casual match on Ancient map:
https://drive.google.com/file/d/1qo3EKnPKP_iwuOW_u4N-mEuUITlsEmi4/view?usp=sharing

I had set the graphic preset ingame to "High" preset. I think this could be VRAM related, as I always experience 100% VRAM usage when this happen (or could be unrelated, as my GPU only has 4GB VRAM (NVIDIA RTX 3050Ti 4GB)).

[danginsburg]

I have taken 3 samples on 3 separate smokes on a Casual match on Ancient map:
https://drive.google.com/file/d/1qo3EKnPKP_iwuOW_u4N-mEuUITlsEmi4/view?usp=sharing

Thank you! Taking a look now!

[danginsburg]

I had set the graphic preset ingame to "High" preset. I think this could be VRAM related, as I always experience 100% VRAM usage when this happen (or could be unrelated, as my GPU only has 4GB VRAM (NVIDIA RTX 3050Ti 4GB)).

If possible could you add -con_logfile to your command line and send us ~/.steam/steam/steamapps/common/Counter-Strike Global Offensive/game/csgo/console.log. You don't need to do that with RenderDoc (or if you have the logs from before, just send those). If there was a memory allocation failure I'd expect something in the log.

[danginsburg]

@huupoke - we have gone through the three RenderDoc captures that you sent, but all of them contain the smoke. To clarify - were you seeing the smokes appear or not when you took these captures? We need captures of the smoke NOT appearing.

This is what I see for example in your capture:

[tiagotriques]

One other detail: would you please post a screenshot of your Video Settings? That might help us repro. Thanks.

It could be VRAM related. That game started at 16:30 and NVidia SMI was reporting only having 23MB of free VRAM at that time

[W33X]

I might try to capture this, the issue appears to happen more often on matches with more players

[huupoke]

To clarify - were you seeing the smokes appear or not when you took these captures? We need captures of the smoke NOT appearing.

@danginsburg I'm certainly sure that the smoke is NOT APPEARING at the time of capture in those 3 captures. In those 3 captures, I was either standing in the smokes and invisible, or the smoke is directly in front of me and invisible.

In tiagotriques's graph above, it seems that VRAM usage drops whenever the capture is taking, so I suspect that's why my capture is still showing the smoke properly?

[tiagotriques]

That graph was during a "normal" gameplay, without the renderdoc capture (:

[huupoke]

@danginsburg
This is my new capture on Inferno:
https://drive.google.com/file/d/1Vcdr75NU6YAMn_jJCXXc3fRyxYoEysT4/view?usp=sharing

My console log:
console.log

Video evidence (taken with Steam recording):
https://drive.google.com/file/d/1bphrWgeoWJ849HglnzZNX4xyv7pdjDD7/view?usp=sharing

[danginsburg]

Thanks! I will take a look at this capture tomorrow. I definitely believe you that the smoke was not appearing in the previous captures. It's interesting that the RenderDoc replay is not reproducing it for us. Did you happen to open it in qrenderdoc and see if on your machine the smoke was visible in the final swapchain image? I'm hoping it isn't, in which case perhaps I just need to open the capture on more similar hardware/driver (I was on a GTX 2080 on a different driver than you).

[huupoke]

Unfortunately, the smoke is visible when I open it in qrenderdoc:

[danginsburg]

The January 28, 2025 CS2 update contained a speculative fix for this bug. We have been unable to reproduce it here, but have been talking to NVIDIA about it. Can you try the new update and report back if you are still seeing the bug? Also, there is no need to gather more captures with RenderDoc - at this point that unfortunately isn't helping since we are unable to reproduce it in the replay.

[tiagotriques]

Apparently the bug is fixed for me. It always happened on my machine when the free VRAM got low, round 10MB.
Just played 3 games, CS2 ate all my VRAM (hehe) but the smokes didn't disappear.

[W33X]

Same here, played couple matches today and got no disappearing smokes

[huupoke]

I think it fixed the bug. I have played about 10 matches since then and haven't seen the bug again.

[W33X]

Sooo, do we have any idea what caused it in the first place?

[danginsburg]

Sooo, do we have any idea what caused it in the first place?

It appears to have been a driver bug with NVIDIA Linux driver implementation of pageable device local memory/memory priority. We're still talking to NVIDIA about the problem, but since we worked around it I will close out this issue. Thanks.