Problem building deviceQuery with Ubuntu 26.04 package cuda-toolkit-13-1

[pubfnbar]

I'm on Ubuntu 26.04 LTS, with CUDA Toolkit 13.1.1 installed via sudo apt install cuda-toolkit-13-1, coming from the Ubuntu packages.

Steps to reproduce:

cd ~/Downloads
git clone https://github.com/NVIDIA/cuda-samples.git
cd ~/Downloads/cuda-samples/cpp/1_Utilities/deviceQuery
mkdir build && cd build
cmake ..

The output:

-- The C compiler identification is GNU 15.2.0
-- The CXX compiler identification is GNU 15.2.0
CMake Error at /usr/share/cmake-4.2/Modules/CMakeDetermineCompilerId.cmake:928 (message):
  Compiling the CUDA compiler identification source file
  "CMakeCUDACompilerId.cu" failed.

  Compiler: /usr/local/cuda-13.1/bin/nvcc

  Build flags:

  Id flags: --keep;--keep-dir;tmp -v



  The output was:

  2

  #$ _NVVM_BRANCH_=nvvm

  #$ _SPACE_=

  #$ _CUDART_=cudart

  #$ _HERE_=/usr/local/cuda-13.1/bin

  #$ _THERE_=/usr/local/cuda-13.1/bin

  #$ _TARGET_SIZE_=

  #$ _TARGET_DIR_=

  #$ _TARGET_DIR_=targets/x86_64-linux

  #$ TOP=/usr/local/cuda-13.1/bin/..

  #$ CICC_PATH=/usr/local/cuda-13.1/bin/../nvvm/bin

  #$ NVVMIR_LIBRARY_DIR=/usr/local/cuda-13.1/bin/../nvvm/libdevice

  #$ LD_LIBRARY_PATH=/usr/local/cuda-13.1/bin/../lib:

  #$
  PATH=/usr/local/cuda-13.1/bin/../nvvm/bin:/usr/local/cuda-13.1/bin:/home/myname/.cargo/bin:/usr/local/cuda-13.1/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin

  #$ INCLUDES="-I/usr/local/cuda-13.1/bin/../targets/x86_64-linux/include"

  #$ SYSTEM_INCLUDES="-isystem"
  "/usr/local/cuda-13.1/bin/../targets/x86_64-linux/include/cccl"

  #$ LIBRARIES=
  "-L/usr/local/cuda-13.1/bin/../targets/x86_64-linux/lib/stubs"
  "-L/usr/local/cuda-13.1/bin/../targets/x86_64-linux/lib"

  #$ CUDAFE_FLAGS=

  #$ PTXAS_FLAGS=

  #$ rm tmp/a_dlink.reg.c

  #$ gcc -D__CUDA_ARCH_LIST__=750 -E -x c++ -D__CUDACC__ -D__NVCC__
  "-I/usr/local/cuda-13.1/bin/../targets/x86_64-linux/include" "-isystem"
  "/usr/local/cuda-13.1/bin/../targets/x86_64-linux/include/cccl"
  -D__CUDACC_VER_MAJOR__=13 -D__CUDACC_VER_MINOR__=1
  -D__CUDACC_VER_BUILD__=115 -D__CUDA_API_VER_MAJOR__=13
  -D__CUDA_API_VER_MINOR__=1 -D__NVCC_DIAG_PRAGMA_SUPPORT__=1
  -D__CUDACC_DEVICE_ATOMIC_BUILTINS__=1 -include "cuda_runtime.h" -m64
  "CMakeCUDACompilerId.cu" -o "tmp/CMakeCUDACompilerId.cpp4.ii"

  #$ cudafe++ --c++17 --static-host-stub --device-hidden-visibility
  --gnu_version=150200 --display_error_number --orig_src_file_name
  "CMakeCUDACompilerId.cu" --orig_src_path_name
  "/home/myname/Downloads/cuda-samples/cpp/1_Utilities/deviceQuery/build/CMakeFiles/4.2.3/CompilerIdCUDA/CMakeCUDACompilerId.cu"
  --allow_managed --m64 --parse_templates --gen_c_file_name
  "tmp/CMakeCUDACompilerId.cudafe1.cpp" --stub_file_name
  "CMakeCUDACompilerId.cudafe1.stub.c" --gen_module_id_file
  --module_id_file_name "tmp/CMakeCUDACompilerId.module_id"
  "tmp/CMakeCUDACompilerId.cpp4.ii"

  /usr/include/x86_64-linux-gnu/bits/mathcalls.h(206): error: exception
  specification is incompatible with that of previous function "rsqrt"
  (declared at line 629 of
  /usr/local/cuda-13.1/bin/../targets/x86_64-linux/include/crt/math_functions.h)

     extern double rsqrt (double __x) noexcept (true); extern double __rsqrt (double __x) noexcept (true);
                                      ^



  /usr/include/x86_64-linux-gnu/bits/mathcalls.h(206): error: exception
  specification is incompatible with that of previous function "rsqrtf"
  (declared at line 653 of
  /usr/local/cuda-13.1/bin/../targets/x86_64-linux/include/crt/math_functions.h)

     extern float rsqrtf (float __x) noexcept (true); extern float __rsqrtf (float __x) noexcept (true);
                                     ^



  2 errors detected in the compilation of "CMakeCUDACompilerId.cu".

  # --error 0x2 --





Call Stack (most recent call first):
  /usr/share/cmake-4.2/Modules/CMakeDetermineCompilerId.cmake:8 (CMAKE_DETERMINE_COMPILER_ID_BUILD)
  /usr/share/cmake-4.2/Modules/CMakeDetermineCompilerId.cmake:53 (__determine_compiler_id_test)
  /usr/share/cmake-4.2/Modules/CMakeDetermineCUDACompiler.cmake:163 (CMAKE_DETERMINE_COMPILER_ID)
  CMakeLists.txt:5 (project)


-- Configuring incomplete, errors occurred!

[philmcm]

1. If you've been playing with the configuration, try deleting and recreating the build directory.
2. Perhaps try changing the cmake .. command to the following to explicitly identify the host compiler to CUDA:

cmake -D CMAKE_CUDA_HOST_COMPILER=g++ ..

Let me know if it works, I'm curious too.

[pubfnbar]

I can't easily test since I already uninstalled the cuda-toolkit-13-1 from Ubuntu 26.04 packages, and installed CUDA 13.3 from NVIDIA network repository, and it works fine. But I found this NVIDIA forum post and this Reddit post claiming that GCC 15.2.0 is incompatible with CUDA 13.1 due to differences in rsqrt and three other related functions.