I’m using Ubuntu 22.04 without a dedicated graphics card. I’ve successfully installed Vapor, but when I try to import vapor, I encounter this error message:
Vapor 3.9.2
Python 3.9.18 (/home/vapor/anaconda3/envs/vapor_python)
[GLContextProviderNvidia.cpp:60] Nvidia GPU not found
*** Break *** segmentation violation
*** Break *** segmentation violation
So, is an Nvidia GPU required?
The code finds an OpenGL context with the help from the OS, and it seems in your case, the code thinks there’s an Nvidia graphics card, but there isn’t.
An OpenGL context is straightforward when there’s a dedicated graphics card; however, when there isn’t, the OS does a wide range of things to try to provide an OpenGL context to applications, and it often works with light workloads, but doesn’t work with heavy workloads. VAPOR definitely falls in the heavy workload case.
So to answer your question, a dedicated GPU is strongly recommended, though it doesn’t have to be Nvidia. Without one, even VAPOR works in some cases, the performance will suffer and become unbearable.
Thank you for your response.
So, in the same computer with the same system, it’s possible that Vapor GUI can be used but Python cannot, is that correct?
I’m in the process of purchasing a dedicated graphics card. If I encounter any other issues, I’ll come back to ask. Thank you very much.
VAPOR GUI also needs to request an OpenGL context from the OS, and it is possible that the GUI coordinates with the OS a little better than Python, thus you don’t see an error anymore. So your best bet is to give it a try. Another aspect of it is performance, where you might use the software just fine on a small dataset, but experience dramatic slowdown with a large dataset. Better to give it a spin and see if it satisfies your workload.