x11 wgpuSurfaceConfigure: Validation Error, Not enough memory left

[FlynnOConnell]

Related: #688

Hi Almar! It was nice to quasi-meet you today! I'd love to sit down for a chat about WGPU/Pygfx soon as I'm starting to use and learn more about it, I think it's incredibly impressive work!

I'm trying to roll a fastplotlib widget out to users on older Ubuntu systems. For some this is the only machine they have access to and it's heavily used, so before I ask the admin to install or update drivers, I really want to make sure I know exactly what the problem is and what needs to be done.

This widget is working on jupyter, and offscreen tests with Vulkan are working, so it seems to me specifically a problem with configuring the canvas during window creation.

Reproducible example

I am testing with two scripts, one to enumerate devices and make an image widget, another to attempt creating a Qt canvas because I have other Qt applications that open on the sever (e.g. the suite2p gui and napari).

1. Test surface creation

import os

# The result is the same no matter what values are entered here. Vulkan worked offscreen
#os.environ['WGPU_BACKEND_TYPE'] = 'Vulkan'
#os.environ['RENDERCANVAS_BACKEND'] = 'qt'

import numpy as np
import wgpu
import fastplotlib as fpl

adapters = wgpu.gpu.enumerate_adapters_sync()
print(f"Found {len(adapters)} adapters\n")

for idx, adapter in enumerate(adapters):
    info = adapter.info
    print(f"[{idx}] {info.get('device')} ({info.get('backend_type')})")
    try:
        device = adapter.request_device_sync()
        limits = device.limits
        print(f"    max-texture-2d: {limits.get('max-texture-dimension-2d')}")
    except Exception as e:
        print(f"    ERROR: {e}")

data = np.random.rand(10, 512, 512).astype(np.float32)
iw = fpl.ImageWidget(data=data, histogram_widget=True)
iw.show()
fpl.loop.run()

Result:

foconnell@v-gpu2:~/repos/mbo_utilities$ RUST_BACKTRACE=full uv run tests/test_surface_creation.py 
Unable to find extension: VK_EXT_physical_device_drm
libEGL warning: egl: failed to create dri2 screen
libEGL warning: DRI2: failed to get drm magic
Found 6 adapters

[0] NVIDIA TITAN V (Vulkan)
    max-texture-2d: 32768
[1] llvmpipe (LLVM 15.0.7, 256 bits) (Vulkan)
    max-texture-2d: 16384
[2] NVIDIA GeForce RTX 3090 Ti (Vulkan)
    max-texture-2d: 32768
[3] NVIDIA TITAN V (Vulkan)
    max-texture-2d: 32768
[4] NVIDIA TITAN V (Vulkan)
    max-texture-2d: 32768
[5] llvmpipe (LLVM 15.0.7, 256 bits) (OpenGL)
    max-texture-2d: 16384

Testing ImageWidget...

thread '<unnamed>' panicked at src/lib.rs:606:5:
Error in wgpuSurfaceConfigure: Validation Error

Caused by:
  Not enough memory left.

stack backtrace:
   0:     0x7f172ad28ba2 - <std::sys::backtrace::BacktraceLock::print::DisplayBacktrace as core::fmt::Display>::fmt::hdcfcb6d4c8489523
   1:     0x7f172ad4a293 - core::fmt::write::h8a494366950f23bb
   2:     0x7f172ad26953 - std::io::Write::write_fmt::h6556609fca33d0b1
   3:     0x7f172ad289f2 - std::sys::backtrace::BacktraceLock::print::hb2a626a81e06b2dc
   4:     0x7f172ad29a93 - std::panicking::default_hook::{{closure}}::h4f78485264f12d10
   5:     0x7f172ad29875 - std::panicking::default_hook::h2c66fc99e962531d
   6:     0x7f172ad2a4a5 - std::panicking::rust_panic_with_hook::h33ac55f64bbd807d
   7:     0x7f172ad2a23a - std::panicking::begin_panic_handler::{{closure}}::h30e7cb89678a57fe
   8:     0x7f172ad290a9 - std::sys::backtrace::__rust_end_short_backtrace::hed60f27456c16ced
   9:     0x7f172ad29ecd - __rustc[de2ca18b4c54d5b8]::rust_begin_unwind
  10:     0x7f172ad47c50 - core::panicking::panic_fmt::h62f63d096dd276af
  11:     0x7f172a8539b6 - wgpu_native::handle_error_fatal::h25a3665c7d3e2579
  12:     0x7f172a869679 - wgpuSurfaceConfigure
  13:     0x7f17321dc052 - ffi_call_unix64
  14:     0x7f17321da9bd - ffi_call_int
  15:     0x7f17321dab38 - cffistatic_ffi_call
  16:     0x7f17321c262c - cdata_call
  17:          0x1620bd8 - _PyEval_EvalFrameDefault
  18:          0x1611cd0 - method_vectorcall.llvm.2544469730125712706
  19:          0x162437e - _PyEval_EvalFrameDefault
  20:          0x1611cd0 - method_vectorcall.llvm.2544469730125712706
  21:          0x162437e - _PyEval_EvalFrameDefault
  22:          0x164bb5c - slot_tp_init
  23:          0x16027a3 - type_call
  24:          0x161497a - _PyEval_EvalFrameDefault
  25:          0x164bc6f - slot_tp_init
  26:          0x16027a3 - type_call
  27:          0x1620bd8 - _PyEval_EvalFrameDefault
  28:          0x164bc6f - slot_tp_init
  29:          0x16027a3 - type_call
  30:          0x161497a - _PyEval_EvalFrameDefault
  31:          0x16996e2 - PyEval_EvalCode
  32:          0x16b3e82 - run_mod.llvm.11839558314017990275
  33:          0x17c005d - pyrun_file
  34:          0x17bfdec - _PyRun_SimpleFileObject
  35:          0x17bfcc2 - _PyRun_AnyFileObject
  36:          0x17bfc04 - pymain_run_file_obj
  37:          0x17bfb18 - pymain_run_file
  38:          0x17558fb - Py_RunMain
  39:          0x17dcd3a - pymain_main.llvm.11357162329980987376
  40:          0x17dcb2d - main
  41:     0x7f17b1c29d90 - __libc_start_call_main
                               at ./csu/../sysdeps/nptl/libc_start_call_main.h:58:16
  42:     0x7f17b1c29e40 - __libc_start_main_impl
                               at ./csu/../csu/libc-start.c:392:3
  43:          0x17bb1ad - _start
  44:                0x0 - <unknown>
fatal runtime error: failed to initiate panic, error 5, aborting

2. Create Qt surface

import os
import wgpu

adapter = wgpu.gpu.request_adapter_sync(power_preference="high-performance")
print(f"  Request adapter sync: {adapter.info.get('device')}")

device = adapter.request_device_sync()
print("  Device created successfully: \n  {device}")

print("\nCreate Qt canvas")
try:
    from qtpy import QtWidgets
    import sys
    from rendercanvas.qt import QRenderCanvas

    print("Creating app")
    app = QtWidgets.QApplication(sys.argv)
    print("  Qt app created")

    print("Creating canvas")
    canvas = QRenderCanvas(size=(800, 600), title="Test")
    print("  Qt canvas created")

    print("Retrieve wgpu context")
    present_context = canvas.get_wgpu_context()
    print(f"  Context: {present_context}")

    print("\nConfigure context")
    render_texture_format = present_context.get_preferred_format(adapter)
    present_context.configure(device=device, format=render_texture_format)
    print("  Context configured")

except Exception as e:
    print(f"\nFailed with error: {e}")
    import traceback

    traceback.print_exc()

foconnell@v-gpu2:~/repos/mbo_utilities$ RUST_BACKTRACE=full uv run tests/test_minimal_surface.py 
Unable to find extension: VK_EXT_physical_device_drm
libEGL warning: egl: failed to create dri2 screen
libEGL warning: DRI2: failed to get drm magic
  Request adapter sync: NVIDIA TITAN V
  Device created successfully: 
  {device}

Create Qt canvas
Creating app
  Qt app created
Creating canvas
libGL error: glx: failed to create dri3 screen
libGL error: failed to load driver: nouveau
libGL error: failed to get magic
libGL error: failed to load driver: nouveau
  Qt canvas created
Retrieve wgpu context
  Context: <rendercanvas.contexts.WgpuContextToScreen object at 0x7f4674a90fe0>

Configure context

thread '<unnamed>' panicked at src/lib.rs:606:5:
Error in wgpuSurfaceConfigure: Validation Error

Caused by:
  Not enough memory left.

stack backtrace:
   0:     0x7f4674728ba2 - <std::sys::backtrace::BacktraceLock::print::DisplayBacktrace as core::fmt::Display>::fmt::hdcfcb6d4c8489523
   1:     0x7f467474a293 - core::fmt::write::h8a494366950f23bb
   2:     0x7f4674726953 - std::io::Write::write_fmt::h6556609fca33d0b1
   3:     0x7f46747289f2 - std::sys::backtrace::BacktraceLock::print::hb2a626a81e06b2dc
   4:     0x7f4674729a93 - std::panicking::default_hook::{{closure}}::h4f78485264f12d10
   5:     0x7f4674729875 - std::panicking::default_hook::h2c66fc99e962531d
   6:     0x7f467472a4a5 - std::panicking::rust_panic_with_hook::h33ac55f64bbd807d
   7:     0x7f467472a23a - std::panicking::begin_panic_handler::{{closure}}::h30e7cb89678a57fe
   8:     0x7f46747290a9 - std::sys::backtrace::__rust_end_short_backtrace::hed60f27456c16ced
   9:     0x7f4674729ecd - __rustc[de2ca18b4c54d5b8]::rust_begin_unwind
  10:     0x7f4674747c50 - core::panicking::panic_fmt::h62f63d096dd276af
  11:     0x7f46742539b6 - wgpu_native::handle_error_fatal::h25a3665c7d3e2579
  12:     0x7f4674269679 - wgpuSurfaceConfigure
  13:     0x7f46759ae052 - ffi_call_unix64
  14:     0x7f46759ac9bd - ffi_call_int
  15:     0x7f46759acb38 - cffistatic_ffi_call
  16:     0x7f467599462c - cdata_call
  17:          0x1620bd8 - _PyEval_EvalFrameDefault
  18:          0x1611cd0 - method_vectorcall.llvm.2544469730125712706
  19:          0x162437e - _PyEval_EvalFrameDefault
  20:          0x1611cd0 - method_vectorcall.llvm.2544469730125712706
  21:          0x162437e - _PyEval_EvalFrameDefault
  22:          0x16996e2 - PyEval_EvalCode
  23:          0x16b3e82 - run_mod.llvm.11839558314017990275
  24:          0x17c005d - pyrun_file
  25:          0x17bfdec - _PyRun_SimpleFileObject
  26:          0x17bfcc2 - _PyRun_AnyFileObject
  27:          0x17bfc04 - pymain_run_file_obj
  28:          0x17bfb18 - pymain_run_file
  29:          0x17558fb - Py_RunMain
  30:          0x17dcd3a - pymain_main.llvm.11357162329980987376
  31:          0x17dcb2d - main
  32:     0x7f4675a29d90 - __libc_start_call_main
                               at ./csu/../sysdeps/nptl/libc_start_call_main.h:58:16
  33:     0x7f4675a29e40 - __libc_start_main_impl
                               at ./csu/../csu/libc-start.c:392:3
  34:          0x17bb1ad - _start
  35:                0x0 - <unknown>
fatal runtime error: failed to initiate panic, error 5, aborting

I was first getting a generic surface creation threads panic error alone, I then asked the admin of this system to add me to the render and video user groups. Only then I received the memory error above.

Also note: I get the VK_EXT errors, and libEGL errors even when successfully opening napari.

Environment

GPU

foconnell@v-gpu2:~/repos/mbo_utilities$ nvidia-smi
Thu Nov 20 11:29:12 2025       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 545.23.06              Driver Version: 545.23.06    CUDA Version: 12.3     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA TITAN V                 On  | 00000000:18:00.0 Off |                  N/A |
| 28%   41C    P2              29W / 250W |     90MiB / 12288MiB |     17%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
|   1  NVIDIA TITAN V                 On  | 00000000:3B:00.0 Off |                  N/A |
| 28%   39C    P8              25W / 250W |      5MiB / 12288MiB |      0%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
|   2  NVIDIA GeForce RTX 3090 Ti     On  | 00000000:5E:00.0 Off |                  Off |
|  0%   29C    P8              19W / 450W |    127MiB / 24564MiB |     12%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
|   3  NVIDIA TITAN V                 On  | 00000000:86:00.0 Off |                  N/A |
| 28%   37C    P8              25W / 250W |      5MiB / 12288MiB |      0%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
|    0   N/A  N/A   4096913      G   /usr/lib/xorg/Xorg                           85MiB |
|    2   N/A  N/A     42373    C+G   ...epos/beast_imaging/.venv/bin/python      114MiB |
+---------------------------------------------------------------------------------------+

(Relavant sectons of) WGPU Report:

██ system:

             platform:  Linux-6.2.0-36-generic-x86_64-with-glibc2.35
python_implementation:  CPython
               python:  3.12.12

██ versions:

        wgpu:  0.28.1
        cffi:  2.0.0
rendercanvas:  2.4.1
        glfw:  2.10.0
       numpy:  2.3.5
       pygfx:  0.15.3
    pylinalg:  0.6.8
 fastplotlib:  0.5.2

██ wgpu_native_info:

expected_version:  27.0.2.0
     lib_version:  27.0.2.0
        lib_path:  ./resources/libwgpu_native-release.so

Environment Variables

foconnell@v-gpu2:~/repos/mbo_utilities$ env
SHELL=/bin/bash
SESSION_MANAGER=local/v-gpu2:@/tmp/.ICE-unix/4097139,unix/v-gpu2:/tmp/.ICE-unix/4097139
WINDOWID=6291463
QT_ACCESSIBILITY=1
COLORTERM=truecolor
XDG_CONFIG_DIRS=/home/foconnell/.config/kdedefaults:/etc/xdg
LANGUAGE=
SHELL_SESSION_ID=f87a6aa5a9564b70b9fa9e26ebaa9462
GTK_RC_FILES=/etc/gtk/gtkrc:/home/foconnell/.gtkrc:/home/foconnell/.config/gtkrc
XCURSOR_SIZE=24
GTK_MODULES=gail:atk-bridge
PWD=/home/foconnell/repos/mbo_utilities
KRB5CCNAME=FILE:/tmp/krb5cc_343984073_ybsQCD
LOGNAME=foconnell
XDG_SESSION_TYPE=x11
SYSTEMD_EXEC_PID=598673
PULSE_SCRIPT=/etc/xrdp/pulse/default.pa
GTK2_RC_FILES=/etc/gtk-2.0/gtkrc:/home/foconnell/.gtkrc-2.0:/home/foconnell/.config/gtkrc-2.0
HOME=/home/foconnell
LANG=en_US.UTF-8
XDG_CURRENT_DESKTOP=KDE
KONSOLE_DBUS_SERVICE=:1.4365
XRDP_SOCKET_PATH=/run/xrdp/sockdir
KONSOLE_DBUS_SESSION=/Sessions/2
PROFILEHOME=
XRDP_PULSE_SOURCE_SOCKET=xrdp_chansrv_audio_in_socket_17
KONSOLE_VERSION=211203
KDE_SESSION_UID=343984073
XDG_SESSION_CLASS=user
TERM=xterm-256color
USER=foconnell
COLORFGBG=15;0
XRDP_SESSION=1
KDE_SESSION_VERSION=5
DISPLAY=:17.0
XRDP_PULSE_SINK_SOCKET=xrdp_chansrv_audio_out_socket_17
SHLVL=1
XDG_SESSION_ID=c124
LD_LIBRARY_PATH=/usr/lib64:/usr/lib
XDG_RUNTIME_DIR=/run/user/343984073
QT_AUTO_SCREEN_SCALE_FACTOR=0
XCURSOR_THEME=breeze_cursors
XDG_DATA_DIRS=/usr/local/share:/usr/share:/var/lib/snapd/desktop
KDE_FULL_SESSION=true
PATH=/home/foconnell/anaconda3/bin:/home/foconnell/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/usr/local/cuda/bin:/usr/local/cuda/bin:/usr/local/cuda/bin
DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/343984073/bus
KDE_APPLICATIONS_AS_SCOPE=1
UID=343984073
KONSOLE_DBUS_WINDOW=/Windows/1
_=/usr/bin/env

Thanks 😃

[FlynnOConnell]

Double checking the required linux packages:

foconnell@v-gpu2:~/repos/mbo_utilities$   dpkg -l | grep -E 'mesa-vulkan-drivers|xserver-xorg-core|libvulkan1'
ii  libvulkan1:amd64                              1.3.204.1-2                                 amd64        Vulkan loader library
ii  mesa-vulkan-drivers:amd64                     23.2.1-1ubuntu3.1~22.04.3                   amd64        Mesa Vulkan graphics drivers
ii  xserver-xorg-core                             2:21.1.4-2ubuntu1.7~22.04.16                amd64        Xorg X server - core server
ii  xserver-xorg-core-hwe-18.04                   3:14.6                                      amd64        Transitional package for xserver-xorg-core-hwe-18.04

[FlynnOConnell]

Not sure if this helps, vulkan summary and vkcube demo:

foconnell@v-gpu2:~/repos/mbo_utilities$ vulkaninfo --summary
WARNING: [Loader Message] Code 0 : terminator_CreateInstance: Failed to CreateInstance in ICD 1.  Skipping ICD.
==========
VULKANINFO
==========

Vulkan Instance Version: 1.3.204

...some instance info omitted 

Instance Layers: count = 3
--------------------------
VK_LAYER_INTEL_nullhw       INTEL NULL HW                1.1.73   version 1
VK_LAYER_MESA_device_select Linux device selection layer 1.3.211  version 1
VK_LAYER_MESA_overlay       Mesa Overlay layer           1.3.211  version 1

Devices:
========
GPU0:
        apiVersion         = 4206852 (1.3.260)
        driverVersion      = 2286272896 (0x8845c180)
        vendorID           = 0x10de
        deviceID           = 0x1d81
        deviceType         = PHYSICAL_DEVICE_TYPE_DISCRETE_GPU
        deviceName         = NVIDIA TITAN V
        driverID           = DRIVER_ID_NVIDIA_PROPRIETARY
        driverName         = NVIDIA
        driverInfo         = 545.23.06
        conformanceVersion = 1.3.6.0
        deviceUUID         = bd3ba5ff-07c5-7a44-0172-26dde958471b
        driverUUID         = 4dc7d163-33e4-5c19-9065-8e7dbc672099
GPU1:
        apiVersion         = 4206847 (1.3.255)
        driverVersion      = 1 (0x0001)
        vendorID           = 0x10005
        deviceID           = 0x0000
        deviceType         = PHYSICAL_DEVICE_TYPE_CPU
        deviceName         = llvmpipe (LLVM 15.0.7, 256 bits)
        driverID           = DRIVER_ID_MESA_LLVMPIPE
        driverName         = llvmpipe
        driverInfo         = Mesa 23.2.1-1ubuntu3.1~22.04.3 (LLVM 15.0.7)
        conformanceVersion = 1.3.1.1
        deviceUUID         = 6d657361-3233-2e32-2e31-2d3175627500
        driverUUID         = 6c6c766d-7069-7065-5555-494400000000
GPU2:
        apiVersion         = 4206852 (1.3.260)
        driverVersion      = 2286272896 (0x8845c180)
        vendorID           = 0x10de
        deviceID           = 0x2203
        deviceType         = PHYSICAL_DEVICE_TYPE_DISCRETE_GPU
        deviceName         = NVIDIA GeForce RTX 3090 Ti
        driverID           = DRIVER_ID_NVIDIA_PROPRIETARY
        driverName         = NVIDIA
        driverInfo         = 545.23.06
        conformanceVersion = 1.3.6.0
        deviceUUID         = 83e02629-cbbb-c602-14f8-cd7f1a4cbab0
        driverUUID         = 4dc7d163-33e4-5c19-9065-8e7dbc672099
GPU3:
        apiVersion         = 4206852 (1.3.260)
        driverVersion      = 2286272896 (0x8845c180)
        vendorID           = 0x10de
        deviceID           = 0x1d81
        deviceType         = PHYSICAL_DEVICE_TYPE_DISCRETE_GPU
        deviceName         = NVIDIA TITAN V
        driverID           = DRIVER_ID_NVIDIA_PROPRIETARY
        driverName         = NVIDIA
        driverInfo         = 545.23.06
        conformanceVersion = 1.3.6.0
        deviceUUID         = bb94506b-67ab-4fc0-27d4-2c5e6c9f968f
        driverUUID         = 4dc7d163-33e4-5c19-9065-8e7dbc672099
GPU4:
        apiVersion         = 4206852 (1.3.260)
        driverVersion      = 2286272896 (0x8845c180)
        vendorID           = 0x10de
        deviceID           = 0x1d81
        deviceType         = PHYSICAL_DEVICE_TYPE_DISCRETE_GPU
        deviceName         = NVIDIA TITAN V
        driverID           = DRIVER_ID_NVIDIA_PROPRIETARY
        driverName         = NVIDIA
        driverInfo         = 545.23.06
        conformanceVersion = 1.3.6.0
        deviceUUID         = 61a50777-e113-f730-7958-1e736b70cd47
        driverUUID         = 4dc7d163-33e4-5c19-9065-8e7dbc672099
foconnell@v-gpu2:~/repos/mbo_utilities$ vkcube
Selected GPU 0: NVIDIA TITAN V, type: 2
Could not find both graphics and present queues

[kushalkolar]

Are you able to run the triangle example?

[FlynnOConnell]

foconnell@v-gpu2:~/repos/mbo_utilities$ uv run ../wgpu-py/examples/triangle.py
Unable to find extension: VK_EXT_physical_device_drm
libEGL warning: egl: failed to create dri2 screen
libEGL warning: DRI2: failed to get drm magic

thread '<unnamed>' panicked at src/lib.rs:606:5:
Error in wgpuSurfaceConfigure: Validation Error

Caused by:
  Not enough memory left.

note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
fatal runtime error: failed to initiate panic, error 5, aborting

[kushalkolar]

run with RUST_BACKTRACE=1

[FlynnOConnell]

foconnell@v-gpu2:~/repos/mbo_utilities$ RUST_BACKTRACE=1 uv run ../wgpu-py/examples/triangle.py
Unable to find extension: VK_EXT_physical_device_drm
libEGL warning: egl: failed to create dri2 screen
libEGL warning: DRI2: failed to get drm magic

thread '<unnamed>' panicked at src/lib.rs:606:5:
Error in wgpuSurfaceConfigure: Validation Error

Caused by:
  Not enough memory left.

stack backtrace:
   0: __rustc::rust_begin_unwind
   1: core::panicking::panic_fmt
   2: wgpu_native::handle_error_fatal
   3: wgpuSurfaceConfigure
   4: ffi_call_unix64
   5: ffi_call_int
   6: cffistatic_ffi_call
   7: cdata_call
   8: _PyEval_EvalFrameDefault
   9: method_vectorcall.llvm.2544469730125712706
  10: _PyEval_EvalFrameDefault
  11: method_vectorcall.llvm.2544469730125712706
  12: _PyEval_EvalFrameDefault
  13: PyEval_EvalCode
  14: run_mod.llvm.11839558314017990275
  15: pyrun_file
  16: _PyRun_SimpleFileObject
  17: _PyRun_AnyFileObject
  18: pymain_run_file_obj
  19: pymain_run_file
  20: Py_RunMain
  21: pymain_main.llvm.11357162329980987376
  22: main
  23: __libc_start_call_main
             at ./csu/../sysdeps/nptl/libc_start_call_main.h:58:16
  24: __libc_start_main_impl
             at ./csu/../csu/libc-start.c:392:3
  25: _start
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
fatal runtime error: failed to initiate panic, error 5, aborting

[kushalkolar]

Those EGL warnings make me want to install new drivers.

[almarklein]

This widget is working on jupyter, and offscreen tests with Vulkan are working, so it seems to me specifically a problem with configuring the canvas during window creation.

Yes, that's likely, and a known problem on Linux. The solution would be to avoid creating a native surface, but to use an offscreen texture, which is then downloaded to RAM so it can be presented as a plain old bitmap. This is called the bitmap present-method, in rendercanvas.

In the triangle example, try changing to these lines:

from rendercanvas.pyside6 import RenderCanvas, loop  # replace pyqt6/pyqt5 if applicable

canvas = RenderCanvas(size=(640, 480), title="wgpu triangle example", present_method='bitmap')

The EGL errors can likely be ignored. Internally, wgpu is enumerating all possible backends, and in that process, some backends (OpenGL in this case) can produce warnings. This can happen even when no OpenGL adapter is ever created. I expect that in time this will be fixed upstream.

[kushalkolar]

libegl might be part of the vulkan ICD loader? https://download.nvidia.com/XFree86/Linux-x86_64/580.95.05/README/installedcomponents.html

GLVND vendor implementation libraries for GLX (/usr/lib/libGLX_nvidia.so.0) and EGL (/usr/lib/libEGL_nvidia.so.0); these libraries provide NVIDIA implementations of OpenGL functionality which may be accessed using the GLVND client-facing libraries.

libGLX_nvidia.so.0 and libEGL_nvidia.so.0 can be used as Vulkan ICDs. The Vulkan ICD configuration file included with the driver references libGLX_nvidia.so.0, however in environments where X11 client libraries are not available, libEGL_nvidia.so.0 should be used.

The Vulkan ICD configuration file is installed as /etc/vulkan/icd.d/nvidia_icd.json.

An additional Vulkan layer configuration file is installed as /etc/vulkan/implicit_layer.d/nvidia_layers.json. These layers add functionality to the Vulkan loader.

Do you have a physical screen attached to the GPU and you have a desktop environment running?

[FlynnOConnell]

ok! we're cooking again!

[FlynnOConnell]

Do you have a physical screen attached to the GPU and you have a desktop environment running?

I don't think so but can likely connect one, and yes

[FlynnOConnell]

so the current solution is to image widget kwargs, which works for now until its fixed upstream.

figure_kwargs = {
    "canvas": "pyside6",
    "canvas_kwargs": {"present_method": "bitmap"},
}
iw = fpl.ImageWidget(
    data,
    figure_kwargs=figure_kwargs,
    
)
iw.show()
fpl.loop.run()

[kushalkolar]

so the current solution is to image widget kwargs, which works for now until its fixed upstream.

figure_kwargs = {
    "canvas": "pyside6",
    "canvas_kwargs": {"present_method": "bitmap"},
}
iw = fpl.ImageWidget(
    data,
    figure_kwargs=figure_kwargs,
    
)
iw.show()
fpl.loop.run()

I'm so confused 😂

[Vipitis]

Side note: if you set your backend (graphics API really) via the instance extras instead of the env var, and do that before enumerating devices you should be able to skip the OpenGL warnings.

See https://wgpu-py.readthedocs.io/en/stable/backends.html#wgpu.backends.wgpu_native.set_instance_extras
It might even be faster, although I haven't really tested that. And this is a native only feature, meaning in pyodide this doesn't exist.