Building CUDA Extension

[sunset1995]

Building CUDA Extension

[greeneggsandyaml]

Thanks for the investigation! I'm using gcc 5.4.0, like in the issue that you linked. I will install a higher version of gcc and try again!

[greeneggsandyaml]

Hello, I just wanted to follow up on this and give an update. I have tried with CUDA 11.1, with matching versions of CUDA, PyTorch, and pytorch-scatter. I still get the same error.

[sunset1995]

Hmm strange. My machine is torch==1.10.1+cu111 too and it works well.
Could you please provide more detail about your nvcc version and which version of the DVGO do you use?

Besides, have you do any modification to the c++/cuda codes?
It's strange that the backend functions use OrderedDict to access torch::Tensor.

[greeneggsandyaml]

Thanks for the response!

I am using the main branch of DVGO (I just cloned, installed the dependencies, and tried to run run.py). I have not modified the code at all. Here is the result of nvcc -V:

>>> nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Mon_Oct_12_20:09:46_PDT_2020
Cuda compilation tools, release 11.1, V11.1.105
Build cuda_11.1.TC455_06.29190527_0

I'm using Python 3.9 and an NVIDIA A40. What GPU/Python/versions are you using?

Thanks so much for your help!

[sunset1995]

On my side, my nvcc -V is

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Tue_Sep_15_19:10:02_PDT_2020
Cuda compilation tools, release 11.1, V11.1.74
Build cuda_11.1.TC455_06.29069683_0

I'm using Python 3.9, torch==1.10.1+cu111.
My GPU is RTX 2080Ti with CUDA 11.1.

[sunset1995]

I found a similar issue in other repos.
It seems that gcc version matters too.
My gcc --version is

gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
Copyright (C) 2017 Free Software Foundation, Inc.

[supdhn]

Hi everyone I am having the same issue and I haven't figured it out why ? any help ?

I installed pytorch through this command:

pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu116

gcc version on my computer:

gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
Copyright (C) 2017 Free Software Foundation, Inc.

g++ version on my computer:

g++ (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
Copyright (C) 2017 Free Software Foundation, Inc.

and nvcc version is

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Tue_Mar__8_18:36:24_Pacific_Standard_Time_2022
Cuda compilation tools, release 11.6, V11.6.124
Build cuda_11.6.r11.6/compiler.31057947_0

Any ideas on how to solve this ? I am on a Ubuntu 18.04 with 16gb memory and RTX Tian. Using python 3.7.5.
Any help is appreciated !

P.S.: I have tried with cuda versions 11.7, 11.6 and 11.3 and they all give me the sames errors

thanks in advance

[j93hahn]

Has anyone received a message like #error -- unsupported GNU version! gcc versions later than 8 are not supported!

I have gcc-12.1.0 installed, is this too new? I'm not entirely sure how to install a gcc version like 7.5.0 as it does not seem to be supported on any conda channels.

[RedemptYourself]

did it work? it seems not

[aztsdfhj]

Thanks for the investigation! I'm using gcc 5.4.0, like in the issue that you linked. I will install a higher version of gcc and try again!

Did you finally solve it? I had the same problem as you