Compilation Error on Ubuntu 24.04.1 LTS with 4060 Ti

[HarryPeverell]

Here's a suggested GitHub issue template for your compilation problem:
Hello,

I am encountering an issue while trying to compile the tiny-cuda-nn project. Here are the details of my setup:

• Hardware: NVIDIA GeForce RTX 4060 Ti
• Operating System: Ubuntu 24.04.1 LTS
• CUDA Version: 12.4 (CUDA compilation tools, release 12.4, V12.4.99)
• CMake Version: 3.26.4

When I execute the following command in the tiny-cuda-nn directory:

cmake --build build --config RelWithDebInfo -j

I receive the following error message:

$ cmake --build build --config RelWithDebInfo -j
-- Obtained CUDA architectures automatically from installed GPUs
-- Targeting CUDA architectures: 89
-- {fmt} version: 11.0.2
-- Build type: RelWithDebInfo
-- Configuring done (0.0s)
-- Generating done (0.0s)
-- Build files have been written to: /home/harry/gs-relight/tiny-cuda-nn/build
[ 15%] Built target fmt
[ 20%] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/cutlass_mlp.cu.o
[ 25%] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/fully_fused_mlp.cu.o
/home/harry/gs-relight/tiny-cuda-nn/dependencies/fmt/include/fmt/base.h(454): warning #128-D: loop is not reachable
    for (; n != 0; ++s1, ++s2, --n) {
    ^
          detected during:
            instantiation of "auto fmt::v11::detail::compare(const Char *, const Char *, std::size_t)->int [with Char=char]" at line 573
            instantiation of "auto fmt::v11::basic_string_view<Char>::compare(fmt::v11::basic_string_view<Char>) const->int [with Char=char]" at line 580
            instantiation of class "fmt::v11::basic_string_view<Char> [with Char=char]" at line 2568
            instantiation of "auto fmt::v11::basic_format_args<Context>::get_id(fmt::v11::basic_string_view<Char>) const->int [with Context=fmt::v11::context, Char=char]" at line 2608

Remark: The warnings can be suppressed with "-diag-suppress <warning-number>"

/home/harry/gs-relight/tiny-cuda-nn/dependencies/fmt/include/fmt/base.h(454): warning #128-D: loop is not reachable
    for (; n != 0; ++s1, ++s2, --n) {
    ^
          detected during:
            instantiation of "auto fmt::v11::detail::compare(const Char *, const Char *, std::size_t)->int [with Char=char]" at line 573
            instantiation of "auto fmt::v11::basic_string_view<Char>::compare(fmt::v11::basic_string_view<Char>) const->int [with Char=char]" at line 580
            instantiation of class "fmt::v11::basic_string_view<Char> [with Char=char]" at line 2568
            instantiation of "auto fmt::v11::basic_format_args<Context>::get_id(fmt::v11::basic_string_view<Char>) const->int [with Context=fmt::v11::context, Char=char]" at line 2608

Error limit reached.
100 errors detected in the compilation of "/home/harry/gs-relight/tiny-cuda-nn/src/cutlass_mlp.cu".
Compilation terminated.
gmake[2]: *** [CMakeFiles/tiny-cuda-nn.dir/build.make:212: CMakeFiles/tiny-cuda-nn.dir/src/fully_fused_mlp.cu.o] Error 4
gmake[2]: *** Waiting for unfinished jobs....
gmake[2]: *** [CMakeFiles/tiny-cuda-nn.dir/build.make:107: CMakeFiles/tiny-cuda-nn.dir/src/cutlass_mlp.cu.o] Error 4
gmake[1]: *** [CMakeFiles/Makefile2:136: CMakeFiles/tiny-cuda-nn.dir/all] Error 2
gmake: *** [Makefile:136: all] Error 2

I would appreciate any guidance or suggestions on how to resolve this issue. Thank you for your help!

Best regards,
Harry

[DeeraDigiSpoc]

Same Error (Since 2022)

[HarryPeverell]

I solved the issue by reinstalling a CUDA version compatible with WSL, following NVIDIA's instructions for WSL:

wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin
sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/12.6.3/local_installers/cuda-repo-wsl-ubuntu-12-6-local_12.6.3-1_amd64.deb
sudo dpkg -i cuda-repo-wsl-ubuntu-12-6-local_12.6.3-1_amd64.deb
sudo cp /var/cuda-repo-wsl-ubuntu-12-6-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-6

After that, the compilation issues I encountered were resolved normally.

[DeeraDigiSpoc]

I solved the issue by reinstalling a CUDA version compatible with WSL, following NVIDIA's instructions for WSL:

wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin
sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/12.6.3/local_installers/cuda-repo-wsl-ubuntu-12-6-local_12.6.3-1_amd64.deb
sudo dpkg -i cuda-repo-wsl-ubuntu-12-6-local_12.6.3-1_amd64.deb
sudo cp /var/cuda-repo-wsl-ubuntu-12-6-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-6

After that, the compilation issues I encountered were resolved normally.

The above mentioned are general installation CUDA Toolkit installation instructions.

The issue Compilation Error:
It occurs while building CMake (NVlabs).