Google Chrome extremely slow to start on nvidia gpu laptop

General system Newbie
1

Hello !

When I start google chrome on my endeavour os laptop with nvidia card, google chrome is extremely slow to start.

I have KDE plasma & wayland environement.

Here is my CPU/GPU config :
CPU: Intel i9-14900HX (32) @ 5.600GHz
GPU: NVIDIA GeForce RTX 4080 Max-Q / Mobile
GPU: Intel Raptor Lake-S UHD Graphics

Nvidia driver version is : 570.133.07

Here is the logs from google-chrome-stable when starting from the command line :

Created TensorFlow Lite XNNPACK delegate for CPU.
[122877:122903:0419/214928.421635:ERROR:registration_request.cc(291)] Registration response error message: DEPRECATED_ENDPOINT
[122877:122903:0419/214948.787934:ERROR:registration_request.cc(291)] Registration response error message: DEPRECATED_ENDPOINT
[0419/214956.193505:ERROR:directory_reader_posix.cc(43)] opendir /home/leoc/.config/google-chrome/Crash Reports/attachments/6f4a40df-e03c-497f-b120-8904409a35bf: No such file or directory (2)
[123011:10:0419/214956.199260:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[123176:1:0419/214956.199258:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[122985:10:0419/214956.199337:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[122994:10:0419/214956.199296:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[122954:1:0419/214956.199267:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[122877:122877:0419/214956.844944:ERROR:gpu_process_host.cc(954)] GPU process exited unexpectedly: exit_code=512
Attempting to use a delegate that only supports static-sized tensors with a graph that has dynamic-sized tensors (tensor#58 is a dynamic-sized tensor).
[122985:10:0419/215026.968408:ERROR:command_buffer_proxy_impl.cc(327)] GPU state invalid after WaitForGetOffsetInRange.
[123176:1:0419/215026.968272:ERROR:command_buffer_proxy_impl.cc(327)] GPU state invalid after WaitForGetOffsetInRange.
[122994:10:0419/215026.968341:ERROR:command_buffer_proxy_impl.cc(327)] GPU state invalid after WaitForGetOffsetInRange.
[122877:122877:0419/215027.613765:ERROR:gpu_process_host.cc(954)] GPU process exited unexpectedly: exit_code=512
[122877:122903:0419/215033.755861:ERROR:registration_request.cc(291)] Registration response error message: DEPRECATED_ENDPOINT
[122954:1:0419/215057.748728:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[122994:10:0419/215057.748876:ERROR:command_buffer_proxy_impl.cc(327)] GPU state invalid after WaitForGetOffsetInRange.
[122877:122877:0419/215058.388889:ERROR:gpu_process_host.cc(954)] GPU process exited unexpectedly: exit_code=512
[122994:10:0419/215058.469659:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[123594:10:0419/215058.469866:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[122954:7:0419/215058.471334:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[123698:123698:0419/215058.509695:ERROR:gl_utils.cc(431)] [.WebGL-0x3cc4000e5c00]GL Driver Message (OpenGL, Performance, GL_CLOSE_PATH_NV, High): GPU stall due to ReadPixels
[123698:123698:0419/215058.531522:ERROR:gl_utils.cc(431)] [.WebGL-0x3cc4000e8d00]GL Driver Message (OpenGL, Performance, GL_CLOSE_PATH_NV, High): GPU stall due to ReadPixels
[123698:123698:0419/215058.544006:ERROR:gl_utils.cc(431)] [.WebGL-0x3cc4000e5c00]GL Driver Message (OpenGL, Performance, GL_CLOSE_PATH_NV, High): GPU stall due to ReadPixels
[123698:123698:0419/215058.558327:ERROR:gl_utils.cc(431)] [.WebGL-0x3cc4000e8d00]GL Driver Message (OpenGL, Performance, GL_CLOSE_PATH_NV, High): GPU stall due to ReadPixels (this message will no longer repeat)

In google chrome, ozone platform is set to wayland.

Do you have any idea where it might come from ?

2

Verboten

3

This isn’t helpful at all. If you don’t have anything helpful to say… stay silent instead of welcoming a new forum member that way.

Are you fiddling around with tensor flow lite within a ipython notebook or such ?
Any fancy chrome extensions being installed which are using tensor flow lite ?
Tensor flow isn’t really part of chrome itself. Thus, you could try to start chrome without any extensions to check which one is causing the trouble.

Start it from the terminal with chrome --disable-extensions

You could also try out ungoogled chromium instead. Or one of the other forks.

Additionally, try this. Should work for chrome as well, if I’m not mistaken.

1 Like
4

I found out that using hybrid mode with envycontrol fixed the issue

1 Like
5

So, you’ve found the solution. For future references, you can mark your reply as the solution. In case someone else stumbles upon the same issue.

Welcome to the forum !

1 Like