Vulkan error during sapien.SapienRender()

[Plankson]

Hi! Thanks for your great work!
I am trying to build sapien-based maniskill simpler_env benchmark in Cloud Container of four Nvidia A800 GPUs.
I've tried many methods today, but the Vulkan driver is still reporting the error:

Before that, I started with:

I have read the official doc and related issue.
What I have done is:

1. when starting the cloud container, I set the NVIDIA_DRIVER_CAPABILITIES environment variable to graphics,utility,compute,display.
2. Inside the container, I installed the packages with apt install libglvnd-dev libegl1 libxext6.
3. I also manually set the VK_ICD_FILENAMES environment variable to /usr/share/vulkan/icd.d/nvidia_icd.json. I notice that in my container, there is no /usr/share/vulkan/icd.d/nvidia_icd.json, and I cannot find this file. So I manually create this file with:

{
    "file_format_version" : "1.0.0",
    "ICD": {
        "library_path": "libGLX_nvidia.so.0",
        "api_version" : "1.2.155"
    }
}

and the same situation with the other /usr/share/glvnd/egl_vendor.d/10_nvidia.json, /etc/vulkan/implicit_layer.d/nvidia_layers.json files mentioned in the official doc. I guess it may have some problem but not sure : (

Although the above steps eliminated the previous error RuntimeError: Cannot find a suitable rendering device, the new error above No Vulkan extensions found for window surface creation persists.

currently the vulkaninfo | grep 'GPU id' output is :

(simplerenv) root@app-bce24d882050496f83d18d49eccefdd7-6797cf4fbc-8x596:/data/users# vulkaninfo | grep 'GPU id'
WARNING: [Loader Message] Code 0 : loader_scanned_icd_add: Driver /usr/lib/x86_64-linux-gnu/libvulkan_intel.so supports Vulkan 1.2, but only supports loader interface version 4. Interface version 5 or newer required to support this version of Vulkan (Policy #LDP_DRIVER_7)
WARNING: [Loader Message] Code 0 : loader_scanned_icd_add: Driver /usr/lib/x86_64-linux-gnu/libvulkan_lvp.so supports Vulkan 1.1, but only supports loader interface version 4. Interface version 5 or newer required to support this version of Vulkan (Policy #LDP_DRIVER_7)
WARNING: [Loader Message] Code 0 : loader_scanned_icd_add: Driver /usr/lib/x86_64-linux-gnu/libvulkan_radeon.so supports Vulkan 1.2, but only supports loader interface version 4. Interface version 5 or newer required to support this version of Vulkan (Policy #LDP_DRIVER_7)
ERROR: [Loader Message] Code 0 : libnvidia-gpucomp.so.550.90.07: cannot open shared object file: No such file or directory
ERROR: [Loader Message] Code 0 : loader_icd_scan: Failed loading library associated with ICD JSON libGLX_nvidia.so.0. Ignoring this JSON
WARNING: [Loader Message] Code 0 : Layer VK_LAYER_MESA_device_select uses API version 1.2 which is older than the application specified API version of 1.4. May cause issues.
'DISPLAY' environment variable not set... skipping surface info
error: XDG_RUNTIME_DIR is invalid or not set in the environment.
error: XDG_RUNTIME_DIR is invalid or not set in the environment.
WARNING: lavapipe is not a conformant vulkan implementation, testing use only.
                GPU id = 0 (llvmpipe (LLVM 12.0.0, 256 bits))
                GPU id = 0 (llvmpipe (LLVM 12.0.0, 256 bits))

What should I do next? I am looking forward to your reply! : )

[fbxiang]

Your vulkaninfo does not pick up any NVIDIA GPU. It seems you do not have libGLX_nvidia.so, which is the core of NVIDIA's graphics driver. You should first confirm libGLX_nvidia.so is indeed missing.

Since you have added graphics capability to the Docker image, libGLX_nvidia.so (and other related libraries) should be mapped from the host machine into the container. So I would guess your host machine is also missing this crucial library (typically due to a partial driver installation). If that is the case, you should reinstall NVIDIA driver on your host machine.