Comfy-Org/ComfyUI_Nvidia_VFX_Nodes
1
0
Fork 0
mirror of https://github.com/Comfy-Org/ComfyUI_Nvidia_VFX_Nodes.git synced 2026-11-04 23:37:30 +00:00
Code Issues 31 Projects Packages Wiki Activity

[GH-ISSUE #20] Cant run inside wsl2 with 5060 ti 16gb is this only for w11? #17

New issue
Open
opened 2026-10-26 22:28:01 +00:00 by bsakaguchi · 7 comments
bsakaguchi commented 2026-10-26 22:28:01 +00:00
Owner
Copy link

Originally created by @alludus on GitHub (Mar 17, 2026).
Original GitHub issue: https://github.com/Comfy-Org/ComfyUI_Nvidia_VFX_Nodes/issues/20

Here's a complete handoff summary:

NVIDIA RTX Video Super Resolution (nvvfx) on WSL2 — Debug Session Summary

Goal: Get the comfyui_nvidia_rtx_nodes ComfyUI custom node working on WSL2 Ubuntu with RTX 5060 Ti 16GB.

Environment:

  • Windows 11 + WSL2 Ubuntu (kernel 6.6.87.2-microsoft-standard-WSL2)
  • GPU: NVIDIA GeForce RTX 5060 Ti 16GB (+ RTX 3060 12GB secondary)
  • Windows Driver: 595.79 (upgraded from 591.74 during session)
  • WSL2 Linux Driver: 570.195.03
  • Python 3.10.18 conda env named comfyui
  • ComfyUI v0.16.4
  • nvidia-vfx 0.1.0.1 (only version available)
  • Launch script: ~/runcomfyui2.sh

The Node:
https://github.com/Comfy-Org/Nvidia_RTX_Nodes_ComfyUI

  • Imports nvvfx (not nvidia.vfx — the node is correct)
  • Uses nvvfx.VideoSuperRes with comfy_api.latest async extension system
  • Node loads fine at ComfyUI startup (0.0 seconds, no error)
  • Crash happens at execution time when sr.load() is called

What Was Ruled Out:

  • ✅ Package installs correctly under nvvfx namespace
  • ✅ import nvvfx works
  • ✅ VideoSuperRes() object creation works
  • ✅ All .so dependencies resolved once LD_LIBRARY_PATH set to ~/anaconda3/envs/comfyui/lib/python3.10/site-packages/nvvfx/libs
  • ✅ TensorRT libs (libnvinfer.so.10 etc.) are bundled inside the pip package itself
  • ✅ NGX runtime (libnvidia-ngx.so) present in WSL2
  • ✅ VSR feature DLLs (nvsvsr.dll, nvvitvsr.dll) present in WSL2 driver directory
  • ✅ /dev/dxg opens successfully (WSL2 GPU path works)
  • ✅ After driver update to 595.79, CUDA 13.2 available
  • ❌ sr.load() silently calls exit_group(1) — no Python exception ever raised

Root Cause (confirmed via strace -f):

The failure is a driver/pip package version mismatch:

  1. sr.load() spawns an NGX worker process
  2. Worker opens: /usr/lib/wsl/drivers/nvmdsi.inf_amd64_d39f1ab212fcacea/nvcubins.bin (595.79 driver's NGX feature container, 73MB)
  3. Worker seeks to offset 64251072, fails to parse it, calls exit_group(1)

The bundled libnvngxruntime.so was built from:

/dvs/p4/build/sw/rel/gpu_drv/r570/r573_76/drivers/ngx/

It cannot parse the r595-format nvcubins.bin from driver 595.79.

Attempting to swap in the r595 libnvidia-ngx.so.1 from the WSL driver directory loads successfully but then libnvVFXVideoSuperRes.so fails with:

undefined symbol: _ZN19NvNGXRuntimeManager11GetInstanceEv

Meaning both libnvngxruntime.so and libnvVFXVideoSuperRes.so are r570-era and the entire VFX SDK in the pip package needs to be rebuilt against the r595 NGX API.

WSL2-Specific Issues Found and Partially Worked Around:

  1. nvidia-modprobe stub required — the SDK spawns it with empty environment (0 vars), bypassing LD_PRELOAD. Real modprobe exits 1 on WSL2 (no kernel module). Fix: replace /usr/bin/nvidia-modprobe with a bash stub that exits 0. Note: gets overwritten on driver updates, must be recreated.

    sudo tee /usr/bin/nvidia-modprobe << 'EOF'
#!/bin/bash
exit 0
EOF
sudo chmod +x /usr/bin/nvidia-modprobe

  2. /usr/share/nvidia/ngx must be created manually — doesn't exist on WSL2:

    sudo mkdir -p /usr/share/nvidia/ngx
sudo chmod 777 /usr/share/nvidia/ngx

  3. /proc/driver/nvidia/capabilities/mig/monitor — doesn't exist on WSL2, appears non-fatal after modprobe stub is in place

  4. LD_LIBRARY_PATH must be set for the bundled .so files — added to runcomfyui2.sh:

    export LD_LIBRARY_PATH=~/anaconda3/envs/comfyui/lib/python3.10/site-packages/nvvfx/libs:$LD_LIBRARY_PATH

Current Status:
Blocked. The nvidia-vfx 0.1.0.1 pip package ships r570-era binaries incompatible with the r595 driver. No newer pip package version exists. Bug reported to NVIDIA Developer Support, case ID: 5986059.

Waiting for: NVIDIA to release an updated nvidia-vfx package built against the r595 NGX SDK.

Next steps when a new package version appears:

  1. pip install --upgrade nvidia-vfx --index-url https://pypi.nvidia.com
  2. Recreate the nvidia-modprobe stub if it was overwritten
  3. Verify LD_LIBRARY_PATH still set in runcomfyui2.sh
  4. Test sr.load() directly before launching ComfyUI
Originally created by @alludus on GitHub (Mar 17, 2026). Original GitHub issue: https://github.com/Comfy-Org/ComfyUI_Nvidia_VFX_Nodes/issues/20 Here's a complete handoff summary: --- **NVIDIA RTX Video Super Resolution (nvvfx) on WSL2 — Debug Session Summary** **Goal:** Get the `comfyui_nvidia_rtx_nodes` ComfyUI custom node working on WSL2 Ubuntu with RTX 5060 Ti 16GB. **Environment:** - Windows 11 + WSL2 Ubuntu (kernel 6.6.87.2-microsoft-standard-WSL2) - GPU: NVIDIA GeForce RTX 5060 Ti 16GB (+ RTX 3060 12GB secondary) - Windows Driver: 595.79 (upgraded from 591.74 during session) - WSL2 Linux Driver: 570.195.03 - Python 3.10.18 conda env named `comfyui` - ComfyUI v0.16.4 - nvidia-vfx 0.1.0.1 (only version available) - Launch script: `~/runcomfyui2.sh` --- **The Node:** `https://github.com/Comfy-Org/Nvidia_RTX_Nodes_ComfyUI` - Imports `nvvfx` (not `nvidia.vfx` — the node is correct) - Uses `nvvfx.VideoSuperRes` with `comfy_api.latest` async extension system - Node loads fine at ComfyUI startup (0.0 seconds, no error) - Crash happens at **execution time** when `sr.load()` is called --- **What Was Ruled Out:** - ✅ Package installs correctly under `nvvfx` namespace - ✅ `import nvvfx` works - ✅ `VideoSuperRes()` object creation works - ✅ All `.so` dependencies resolved once `LD_LIBRARY_PATH` set to `~/anaconda3/envs/comfyui/lib/python3.10/site-packages/nvvfx/libs` - ✅ TensorRT libs (`libnvinfer.so.10` etc.) are bundled inside the pip package itself - ✅ NGX runtime (`libnvidia-ngx.so`) present in WSL2 - ✅ VSR feature DLLs (`nvsvsr.dll`, `nvvitvsr.dll`) present in WSL2 driver directory - ✅ `/dev/dxg` opens successfully (WSL2 GPU path works) - ✅ After driver update to 595.79, CUDA 13.2 available - ❌ `sr.load()` silently calls `exit_group(1)` — no Python exception ever raised --- **Root Cause (confirmed via `strace -f`):** The failure is a **driver/pip package version mismatch**: 1. `sr.load()` spawns an NGX worker process 2. Worker opens: `/usr/lib/wsl/drivers/nvmdsi.inf_amd64_d39f1ab212fcacea/nvcubins.bin` (595.79 driver's NGX feature container, 73MB) 3. Worker seeks to offset 64251072, fails to parse it, calls `exit_group(1)` The bundled `libnvngxruntime.so` was built from: ``` /dvs/p4/build/sw/rel/gpu_drv/r570/r573_76/drivers/ngx/ ``` It **cannot parse the r595-format `nvcubins.bin`** from driver 595.79. Attempting to swap in the r595 `libnvidia-ngx.so.1` from the WSL driver directory loads successfully but then `libnvVFXVideoSuperRes.so` fails with: ``` undefined symbol: _ZN19NvNGXRuntimeManager11GetInstanceEv ``` Meaning **both** `libnvngxruntime.so` and `libnvVFXVideoSuperRes.so` are r570-era and the entire VFX SDK in the pip package needs to be rebuilt against the r595 NGX API. --- **WSL2-Specific Issues Found and Partially Worked Around:** 1. **`nvidia-modprobe` stub required** — the SDK spawns it with empty environment (0 vars), bypassing LD_PRELOAD. Real modprobe exits 1 on WSL2 (no kernel module). Fix: replace `/usr/bin/nvidia-modprobe` with a bash stub that exits 0. Note: gets overwritten on driver updates, must be recreated. ```bash sudo tee /usr/bin/nvidia-modprobe << 'EOF' #!/bin/bash exit 0 EOF sudo chmod +x /usr/bin/nvidia-modprobe ``` 2. **`/usr/share/nvidia/ngx` must be created manually** — doesn't exist on WSL2: ```bash sudo mkdir -p /usr/share/nvidia/ngx sudo chmod 777 /usr/share/nvidia/ngx ``` 3. **`/proc/driver/nvidia/capabilities/mig/monitor`** — doesn't exist on WSL2, appears non-fatal after modprobe stub is in place 4. **`LD_LIBRARY_PATH` must be set** for the bundled `.so` files — added to `runcomfyui2.sh`: ```bash export LD_LIBRARY_PATH=~/anaconda3/envs/comfyui/lib/python3.10/site-packages/nvvfx/libs:$LD_LIBRARY_PATH ``` --- **Current Status:** Blocked. The `nvidia-vfx 0.1.0.1` pip package ships r570-era binaries incompatible with the r595 driver. No newer pip package version exists. Bug reported to NVIDIA Developer Support, **case ID: 5986059**. **Waiting for:** NVIDIA to release an updated `nvidia-vfx` package built against the r595 NGX SDK. **Next steps when a new package version appears:** 1. `pip install --upgrade nvidia-vfx --index-url https://pypi.nvidia.com` 2. Recreate the `nvidia-modprobe` stub if it was overwritten 3. Verify `LD_LIBRARY_PATH` still set in `runcomfyui2.sh` 4. Test `sr.load()` directly before launching ComfyUI
bsakaguchi commented 2026-10-26 22:28:02 +00:00
Author
Owner
Copy link

@doogyhatts commented on GitHub (Mar 22, 2026):

I see. That is why I had a crash.

<!-- gh-comment-id:4105157436 --> @doogyhatts commented on GitHub (Mar 22, 2026): I see. That is why I had a crash.
bsakaguchi commented 2026-10-26 22:28:02 +00:00
Author
Owner
Copy link

@aiberdonml commented on GitHub (Aug 5, 2026):

The failure was resolved by setting the VFX and WSL driver library paths before Python starts

<!-- gh-comment-id:5189482837 --> @aiberdonml commented on GitHub (Aug 5, 2026): The failure was resolved by setting the VFX and WSL driver library paths before Python starts
bsakaguchi commented 2026-10-26 22:28:03 +00:00
Author
Owner
Copy link

@Dunc4n1dah0 commented on GitHub (Aug 10, 2026):

The failure was resolved by setting the VFX and WSL driver library paths before Python starts

I think I have same kind of error - as soon as I try to run the node - Comfy just exits. No error in terminal, nothing.
Can you share your magic line?

<!-- gh-comment-id:5234870881 --> @Dunc4n1dah0 commented on GitHub (Aug 10, 2026): > The failure was resolved by setting the VFX and WSL driver library paths before Python starts I think I have same kind of error - as soon as I try to run the node - Comfy just exits. No error in terminal, nothing. Can you share your magic line?
bsakaguchi commented 2026-10-26 22:28:03 +00:00
Author
Owner
Copy link

@FaintySmile commented on GitHub (Aug 10, 2026):

I fixed the NvVFX_Load failed: The effect has not been properly initialized (code -12) error on Ubuntu 26.04 with an RTX 5060 Ti.

My setup:

GPU: NVIDIA GeForce RTX 5060 Ti
Driver: 595.71.05
Ubuntu: 26.04 LTS
Kernel: 7.0.0-29-generic
PyTorch: 2.13.0+cu132
CUDA: 13.2
nvidia-vfx: 0.1.0.1

The important part was that CUDA itself was working fine:

torch.cuda.is_available() == True

but NvVFX failed at:

sr.load()

with:

nvvfx.NvVFXError:
NvVFX_Load failed:
The effect has not been properly initialized (code -12)

I traced it with strace and found that NvVFX was trying to load:

libnvidia-ngx.so.1

but the library was missing:

/usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1 = ENOENT

For the NVIDIA 595 server driver branch, installing/reinstalling the matching GL package fixed it:

sudo apt update
sudo apt install --reinstall libnvidia-gl-595-server
sudo ldconfig

After that I had:

/usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.595.71.05
/usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1 -> libnvidia-ngx.so.595.71.05

and:

ldconfig -p | grep -i ngx

returned:

libnvidia-ngx.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1

After restarting Python/ComfyUI, NvVFX_Load worked successfully.

For people using the normal non-server 595 driver branch, the equivalent package should be:

sudo apt install --reinstall libnvidia-gl-595
sudo ldconfig

So before downgrading drivers or reinstalling Ubuntu, check:

ls -la /usr/lib/x86_64-linux-gnu/libnvidia-ngx*
ldconfig -p | grep -i ngx

If libnvidia-ngx.so.1 is missing, make sure the matching libnvidia-gl-* package for your installed NVIDIA driver branch is installed.

To see which NVIDIA branch you have:

dpkg -l | grep -E 'nvidia-driver|libnvidia-gl|nvidia-utils'

In my case the relevant installed packages were:

libnvidia-gl-595-server
nvidia-utils-595-server

So the fix was specifically libnvidia-gl-595-server.

This resolved the error without changing Ubuntu, CUDA, PyTorch, or the NVIDIA driver version.

<!-- gh-comment-id:5245720270 --> @FaintySmile commented on GitHub (Aug 10, 2026): I fixed the `NvVFX_Load failed: The effect has not been properly initialized (code -12)` error on Ubuntu 26.04 with an RTX 5060 Ti. My setup: ```text GPU: NVIDIA GeForce RTX 5060 Ti Driver: 595.71.05 Ubuntu: 26.04 LTS Kernel: 7.0.0-29-generic PyTorch: 2.13.0+cu132 CUDA: 13.2 nvidia-vfx: 0.1.0.1 ``` The important part was that CUDA itself was working fine: ```text torch.cuda.is_available() == True ``` but NvVFX failed at: ```python sr.load() ``` with: ```text nvvfx.NvVFXError: NvVFX_Load failed: The effect has not been properly initialized (code -12) ``` I traced it with `strace` and found that NvVFX was trying to load: ```text libnvidia-ngx.so.1 ``` but the library was missing: ```text /usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1 = ENOENT ``` For the NVIDIA 595 **server** driver branch, installing/reinstalling the matching GL package fixed it: ```bash sudo apt update sudo apt install --reinstall libnvidia-gl-595-server sudo ldconfig ``` After that I had: ```text /usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.595.71.05 /usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1 -> libnvidia-ngx.so.595.71.05 ``` and: ```bash ldconfig -p | grep -i ngx ``` returned: ```text libnvidia-ngx.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libnvidia-ngx.so.1 ``` After restarting Python/ComfyUI, `NvVFX_Load` worked successfully. For people using the normal non-server 595 driver branch, the equivalent package should be: ```bash sudo apt install --reinstall libnvidia-gl-595 sudo ldconfig ``` So before downgrading drivers or reinstalling Ubuntu, check: ```bash ls -la /usr/lib/x86_64-linux-gnu/libnvidia-ngx* ldconfig -p | grep -i ngx ``` If `libnvidia-ngx.so.1` is missing, make sure the matching `libnvidia-gl-*` package for your installed NVIDIA driver branch is installed. To see which NVIDIA branch you have: ```bash dpkg -l | grep -E 'nvidia-driver|libnvidia-gl|nvidia-utils' ``` In my case the relevant installed packages were: ```text libnvidia-gl-595-server nvidia-utils-595-server ``` So the fix was specifically `libnvidia-gl-595-server`. This resolved the error without changing Ubuntu, CUDA, PyTorch, or the NVIDIA driver version.
bsakaguchi commented 2026-10-26 22:28:03 +00:00
Author
Owner
Copy link

@Dunc4n1dah0 commented on GitHub (Aug 10, 2026):

My solution; since I made standalone upscaler with webui and it died just like Comfy.

export LD_LIBRARY_PATH="/usr/lib/wsl/lib:$LD_LIBRARY_PATH"
that is all I needed before starting it on WSL.

<!-- gh-comment-id:5246207931 --> @Dunc4n1dah0 commented on GitHub (Aug 10, 2026): My solution; since I made standalone upscaler with webui and it died just like Comfy. `export LD_LIBRARY_PATH="/usr/lib/wsl/lib:$LD_LIBRARY_PATH"` that is all I needed before starting it on WSL.
bsakaguchi commented 2026-10-26 22:28:03 +00:00
Author
Owner
Copy link

@wywywywy commented on GitHub (Aug 18, 2026):

My solution; since I made standalone upscaler with webui and it died just like Comfy.

export LD_LIBRARY_PATH="/usr/lib/wsl/lib:$LD_LIBRARY_PATH" that is all I needed before starting it on WSL.

Just wanted to confirm that this above has fixed it for me in WSL2 with v610.x driver. Thank you.

My LD_LIBRARY_PATH now looks like this - /usr/lib/wsl/lib:/usr/local/cuda-13/lib64:

<!-- gh-comment-id:5325738506 --> @wywywywy commented on GitHub (Aug 18, 2026): > My solution; since I made standalone upscaler with webui and it died just like Comfy. > > `export LD_LIBRARY_PATH="/usr/lib/wsl/lib:$LD_LIBRARY_PATH"` that is all I needed before starting it on WSL. Just wanted to confirm that this above has fixed it for me in WSL2 with v610.x driver. Thank you. My `LD_LIBRARY_PATH` now looks like this - `/usr/lib/wsl/lib:/usr/local/cuda-13/lib64:`
bsakaguchi commented 2026-10-26 22:28:03 +00:00
Author
Owner
Copy link

@925-Studio commented on GitHub (Oct 17, 2026):

5. My solution; since I made standalone upscaler with webui and it died just like Comfy.
export LD_LIBRARY_PATH="/usr/lib/wsl/lib:$LD_LIBRARY_PATH"
that is all I needed before starting it on WSL.

Thank you sir. You just made my day.

<!-- gh-comment-id:6230270320 --> @925-Studio commented on GitHub (Oct 17, 2026): > 5\. My solution; since I made standalone upscaler with webui and it died just like Comfy. > `export LD_LIBRARY_PATH="/usr/lib/wsl/lib:$LD_LIBRARY_PATH"` > that is all I needed before starting it on WSL. Thank you sir. You just made my day.
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
alexis/rename-true-hdr-to-video-hdr
Revert-node-name
alexis/add-long-video-true-hdr-workflow
alexis/add-node-thumbnails-readme
alexis/add-video-upscale-thumbnail
alexis/update-workflow-thumbnails
alexis/update-workflows
Update-README
alexis/update-tru-hdr-workflow
alexis/update-vfx-nodes
alexis/update-pyproject
add-publish-workflow-and-pyproject
No results found.
Labels
Clear labels   
pull-request
Mirrored from GitHub Pull Request

No labels
pull-request
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
Comfy-Org/ComfyUI_Nvidia_VFX_Nodes#17
No description provided.