Open haoyuhsu opened 4 months ago
Could you try conda install nvidia/label/cuda-11.8.0::cuda-cupti
?
It seems that you install cuda via conda, in which case some extra CUDA library would not be installed automatically.
The /usr/bin/ld
can not find the cupti. Installing them manually may solve the problem.
In a system level CUDA (not in the conda environment), they can be found in the extras
folder
I tried conda install nvidia/label/cuda-11.8.0::cuda-cupti
and get the messages # All requested packages already installed. So I think the only way might be installing locally.
Maybe this helps. https://discuss.pytorch.org/t/build-errors-undefined-ref-to-cuptixxx-solved/183318
Yeah a sure-fire way is to install a local CUDA. it's not a hassle and doesn't require sudo anyway.
Unfortunately, I installed CUDA locally and still got the same error.
(chatsim) root@server:~/Documents/maxhsu/editing-related-works/ChatSim/chatsim/background/mcnerf$ which nvcc
/home/haoyuyh3/Documents/maxhsu/cuda/cuda-11.8/bin/nvcc
(chatsim) root@server:~/Documents/maxhsu/editing-related-works/ChatSim/chatsim/background/mcnerf$ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Sep_21_10:33:58_PDT_2022
Cuda compilation tools, release 11.8, V11.8.89
Build cuda_11.8.r11.8/compiler.31833905_0
I've also tried the solution above and added those 4 lines into CMakeLists.txt. Still not working.
I observe the only difference would be one of the cmake warnings disappeared.
CMake Warning at CMakeLists.txt:74 (add_executable):
Cannot generate a safe runtime search path for target main because files in
some directories may conflict with libraries in implicit directories:
runtime library [libnvrtc.so.11.2] in /usr/lib/x86_64-linux-gnu may be hidden by files in:
/home/haoyuyh3/Documents/maxhsu/cuda/cuda-11.8/lib64
runtime library [libnvToolsExt.so.1] in /usr/lib/x86_64-linux-gnu may be hidden by files in:
/home/haoyuyh3/Documents/maxhsu/cuda/cuda-11.8/lib64
Some of these libraries may not be found correctly.
You may try to specify those two libraries in the CMakeLists.txt, so that the cmake can find them explicitly.
I've tried that. Unfortunately, it also failed...Orz
sad to hear that, 3dgs is really a good alternative. The installation is much easier
export CMAKE_PREFIX_PATH=$CONDA_PREFIX
It is worth a final try :)
yeah I think I'll choose to move onto 3dgs. still really appreciate your kind and detailed response!
yeah I think I'll choose to move onto 3dgs. still really appreciate your kind and detailed response!
Hi, i'm wondering if you can successfully run on single RTX 4090?
I tested with my 4090, it is okay
I tested with my 4090, it is okay
Thank you so much!
Hi,
I have tried to install ChatSim environment on NVIDIA RTX 4090, and I have encountered some problems during installation.
OS: Ubuntu 22.04.4 LTS gcc: 11.4.0 cmake: 3.22.1 cuda: 11.8
I have modified several steps to make it compatible to 4090. (since it is compute_89)
As I tried to run
cmake . -B build
, it comes with a warning messages.Afterwards, I ran
cmake --build build --target main --config RelWithDebInfo -j
and came across this error.I would like to know which part did I get it wrong and what is the proper way to fix it? Thanks!