Build fail - recipe for target 'testbed' failed

[kajc10]

Build fails on headless container, any ideas what could be the problem? Specs: Ubuntu 18.04.6 LTS CUDA 11.6 Tesla V100 (DGX station 4pcs) Python 3.9.7 cmake 3.22.3 gcc/g++ 7.5.0

root@610d9320bd51:~/workdir/instant-ngp# cmake . -B build -DNGP_BUILD_WITH_GUI=off

-- !!! Warning OptiX_INSTALL_DIR not set in environment. using default
-- OptiX_INSTALL_DIR value: /usr/local/NVIDIA-OptiX-SDK-7.3.0-linux64-x86_64
-- pybind11 v2.7.1
CMake Warning (dev) at /root/workdir/cmake-3.22.3/share/cmake-3.22/Modules/CMakeDependentOption.cmake:84 (message):
  Policy CMP0127 is not set: cmake_dependent_option() supports full Condition
  Syntax.  Run "cmake --help-policy CMP0127" for policy details.  Use the
  cmake_policy command to set the policy and suppress this warning.
Call Stack (most recent call first):
  dependencies/pybind11/CMakeLists.txt:98 (cmake_dependent_option)
This warning is for project developers.  Use -Wno-dev to suppress it.

-- Targeting GPU architectures: 70;70;70;70
CMake Warning at dependencies/tiny-cuda-nn/CMakeLists.txt:112 (message):
  Fully fused MLPs do not support GPU architectures of 70 or less.  Falling
  back to CUTLASS MLPs.  Remove GPU architectures 70 and lower to allow
  maximum performance

root@610d9320bd51:~/workdir/instant-ngp# cmake --build build --config RelWithDebInfo -j 16 ...long output... end of output:

collect2: error: ld returned 1 exit status
CMakeFiles/testbed.dir/build.make:98: recipe for target 'testbed' failed
make[2]: *** [testbed] Error 1
CMakeFiles/Makefile2:192: recipe for target 'CMakeFiles/testbed.dir/all' failed
make[1]: *** [CMakeFiles/testbed.dir/all] Error 2
Makefile:90: recipe for target 'all' failed
make: *** [all] Error 2

The full output is too long, I attach it as a link: error.txt

[leventt]

Do you maybe have multiple CUDA SDK installations on your system?

Can you print what you see when you run: nvcc -v

I would check the build env variables and perhaps set something like this explicitly to rule it out:

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

What may be happening here is that CMAKE may be finding an old CUDA SDK on your system and can not link some of the symbols from newer CUDA versions.

[kajc10]

Yes, I did have some problems with CUDA. (I'm working on a remote container and at every reset my env variables were lost -> nvcc -V showed CUDA 9.1 instead of 11.6. To solve it I only wrote the following lines to .bash-profile) export PATH=/usr/local/cuda-11.6/bin${PATH:+:${PATH}} export LD_LIBRARY_PATH=/usr/local/cuda-11.6/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

Now issuing your snippet, the build completes, thanks.