MAKE_CUDA_ARCHITECTURES must be non-empty if set.

[narviii]

Getting this error when trying to compile it on Ubuntu. Cuda is installed ( v 11.3), did everything according to the readme. What might be wrong?

Full error text:

CMake Error at /usr/share/cmake-3.24/Modules/CMakeDetermineCUDACompiler.cmake:277 (message): CMAKE_CUDA_ARCHITECTURES must be non-empty if set. Call Stack (most recent call first): CMakeLists.txt:11 (project)

-- Configuring incomplete, errors occurred! See also "/home/narvi/apps/instant-ngp/build/CMakeFiles/CMakeOutput.log". See also "/home/narvi/apps/instant-ngp/build/CMakeFiles/CMakeError.log".

[Tom94]

Can you try again after updating to the latest commit?

git pull origin master
git submodule update --recursive

[narviii]

Same error:(

CMake Error at /usr/share/cmake-3.24/Modules/CMakeDetermineCUDACompiler.cmake:277 (message):
  CMAKE_CUDA_ARCHITECTURES must be non-empty if set.
Call Stack (most recent call first):
  CMakeLists.txt:11 (project)


-- Configuring incomplete, errors occurred!
See also "/home/narvi/apps/instant-ngp/build/CMakeFiles/CMakeOutput.log".
See also "/home/narvi/apps/instant-ngp/build/CMakeFiles/CMakeError.log".
(base) narvi@narviBUBUNTU:~/apps/instant-ngp$ 

[XinyueZ]

same

[tywenk]

Also have this same exact error after running make . -B build

I have all prereqs installed.

Ubuntu version:

Distributor ID:	Ubuntu
Description:	Ubuntu 22.10
Release:	22.10
Codename:	kinetic

gcc version: gcc (Ubuntu 12.2.0-3ubuntu1) 12.2.0

cmake: cmake version 3.24.2

nvidia-smi:

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.60.13    Driver Version: 525.60.13    CUDA Version: 12.0     |
|-------------------------------+----------------------+----------------------+
| 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 GeForce ...  On   | 00000000:09:00.0  On |                  N/A |
|  0%   47C    P5    22W / 240W |   1596MiB /  8192MiB |     29%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+

[tom-bu]

I had to delete the previous build, add the cuda paths in my bashrc, rebuild it and it worked.

[Luoyingfeng8]

I had to delete the previous build, add the cuda paths in my bashrc, rebuild it and it worked.

work, thk~

[ghost]

@tom-bu Can you please say exactly what you added to .bashrc?

[difficulToName]

I have encountered this problem before when I was installing COLMAP. To solve it: rm -r /path/to/instant-ngp/build to delete previous build files vim /path/to/instant-ngp/CMakeList.txt add set(CMAKE_CUDA_ARCHITECTURES "native") under the project()code block save txt and recompile for more information

[Fernweh-yang]

two steps to solve the problem:

1. delete the build folder
2. run: cmake . -B build -DCMAKE_CUDA_COMPILER=/usr/local/cuda-12.1/bin/nvcc -DTCNN_CUDA_ARCHITECTURES=61

[xpmemeda]

@tom-bu Can you please say exactly what you added to .bashrc?

Add your nvcc executable to PATH. In my environment, it is export PATH=$PATH:/usr/local/cuda/bin.

[amughrabi]

Using Anaconda/Linux:

I solved my problem as below:

rm -rf build 

conda install -c nvidia cuda-toolkit

cmake . -B build -DCMAKE_CUDA_COMPILER=$(which nvcc)

[nocnestudio]

Using Anaconda/Linux:

I solved my problem as below:

rm -rf build 

conda install -c nvidia cuda-toolkit

cmake . -B build -DCMAKE_CUDA_COMPILER=$(which nvcc)

You are my HERO. You are the best. Thank You Thank You. I'm crying ;;;;;;((((( THX THX

[sundeco]

Sorry for the basic question, but when I run the cmake command it tells me I need a build directory and I need the CMakeList.txt file in it. Am I correct to assume that after deleting the build directory, I need to use mkdir again and then move CMakeList.txt into that directory?

[Nuclear6]

@tom-bu Can you please say exactly what you added to .bashrc?您能具体说明一下您添加到 .bashrc 中的内容吗？

export PATH="/usr/local/cuda/bin:$PATH" export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

[shink]

I deleted the build directory, then it worked.

[d-bharadwaj]

I had to delete the previous build, add the cuda paths in my bashrc, rebuild it and it worked.

This works, thanks so much.

[iamstarlee]

If you compile multiple libraries at the same time, you need to delete the build inside onnxruntime directory. Because in my case, when I deleted build out of onnxruntime directory, nothing changed.