Can't find CUDA_INCLUDE_DIRS in latest CMAKE [duplicate] - cmake

This question already has answers here:
Listing include_directories in CMake
(3 answers)
Closed 3 years ago.
Since in CMAKE 3.10, CUDA macro is supported by default (https://cmake.org/cmake/help/latest/module/FindCUDA.html).
But I can't find the variable CUDA_INCLUDE_DIRS
cmake_minimum_required(VERSION 3.10 FATAL_ERROR)
project(cmake_and_cuda LANGUAGES CXX CUDA)
message(${CUDA_INCLUDE_DIRS})
the errors are
-- The CXX compiler identification is GNU 5.4.0
-- The CUDA compiler identification is NVIDIA 10.0.130
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc
-- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc -- works
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
CMake Error at CMakeLists.txt:7 (message):
message called with incorrect number of arguments
-- Configuring incomplete, errors occurred!
See also "/home/tumh/code-samples/posts/cmake/build/CMakeFiles/CMakeOutput.log".
Any idea?

Just use the ${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES}.
In all .cu src files, there is no need to manually add
include_directories(${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES}),
because CMake will do it for you.
But for .cpp src files, if you need to include <cuda_runtime.h>, you must manually add
include_directories(${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES})
in your CMakeLists.txt.

Related

CGAL cmake on ubuntu 16.04 fails

I am trying to run the code of PU-GAN(https://github.com/liruihui/PU-GAN),you can get the program(evaluation code) in (https://github.com/liruihui/PU-GAN).but I am failed to camake CGAL libaray. i get following error:
(tf) li#li-System-Product-Name:/media/li/1d105677-e036-4fa4-8e37-124cb400f24d/user/shenbin/PU-GAN-master/evaluation_code$ cmake .
-- The C compiler identification is GNU 9.3.0
-- The CXX compiler identification is GNU 5.4.0
-- Check for working C compiler: /home/li/anaconda3/envs/tf/bin/x86_64-conda-linux-gnu-cc
-- Check for working C compiler: /home/li/anaconda3/envs/tf/bin/x86_64-conda-linux-gnu-cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found OpenMP_C: -fopenmp (found version "4.5")
-- Found OpenMP_CXX: -fopenmp (found version "4.0")
-- Found OpenMP: TRUE (found version "4.5")
-- Using header-only CGAL
-- Targetting Unix Makefiles
-- Using /usr/bin/c++ compiler.
-- Found GMP: /home/li/anaconda3/envs/tf/lib/libgmp.so
-- Found MPFR: /home/li/anaconda3/envs/tf/lib/libmpfr.so
-- Found Boost: /usr/include (found suitable version "1.58.0", minimum required is "1.48")
-- Boost include dirs: /usr/include
-- Boost libraries:
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- Using gcc version 4 or later. Adding -frounding-math
-- Build type:
-- USING CXXFLAGS = ' -fopenmp '
-- USING EXEFLAGS = '-Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,-rpath,/home/li/anaconda3/envs/tf/lib -Wl,-rpath-link,/home/li/anaconda3/envs/tf/lib -L/home/li/anaconda3/envs/tf/lib '
CMake Warning at /usr/local/lib/cmake/CGAL/CGAL_enable_end_of_configuration_hook.cmake:99 (message):
=======================================================================
CGAL performance notice:
The variable CMAKE_BUILD_TYPE is set to "". For performance reasons, you
should set CMAKE_BUILD_TYPE to "Release".
Set CGAL_DO_NOT_WARN_ABOUT_CMAKE_BUILD_TYPE to TRUE if you want to disable
this warning.
=======================================================================
Call Stack (most recent call first):
CMakeLists.txt:9999 (CGAL_run_at_the_end_of_configuration)
-- Configuring done
-- Generating done
-- Build files have been written to: /media/li/1d105677-e036-4fa4-8e37-124cb400f24d/user/shenbin/PU-GAN-master/evaluation_code
(tf) li#li-System-Product-Name:/media/li/1d105677-e036-4fa4-8e37-124cb400f24d/user/shenbin/PU-GAN-master/evaluation_code$ make
Scanning dependencies of target evaluation
[ 50%] Building CXX object CMakeFiles/evaluation.dir/evaluation.cpp.o
/media/li/1d105677-e036-4fa4-8e37-124cb400f24d/user/shenbin/PU-GAN-master/evaluation_code/evaluation.cpp:19:50: fatal error: CGAL/Polygon_mesh_processing/measure.h: No such file or directory
compilation terminated.
CMakeFiles/evaluation.dir/build.make:62: recipe for target 'CMakeFiles/evaluation.dir/evaluation.cpp.o' failed
make[2]: *** [CMakeFiles/evaluation.dir/evaluation.cpp.o] Error 1
CMakeFiles/Makefile2:77: recipe for target 'CMakeFiles/evaluation.dir/all' failed
make[1]: *** [CMakeFiles/evaluation.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2
I cannot find a way to compile any of my CGAL Programyou
It's probably a mismatch between compilers or mismatch between the CGAL library and compilers - your CMake has found the cc compiler inside your anaconda3 directory:
-- The C compiler identification is GNU 9.3.0
-- The CXX compiler identification is GNU 5.4.0
-- Check for working C compiler: /home/li/anaconda3/envs/tf/bin/x86_64-conda-linux-gnu-cc
-- Check for working C compiler: /home/li/anaconda3/envs/tf/bin/x86_64-conda-linux-gnu-cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
Different versions of C and CXX - that's not good. Please do not use the old compiler from the Anaconda distribution, and make sure both C and CXX compilers belong to the same (better the most recent) package.

CMake can't find fortran even tho i have it installed

I've been trying to edit cura by ultimaker source code and trying to compile it using CMake but I can't seem to be able to do it even though I have fortran installed
(i use manjaro linux)
here is my CMake console output
The C compiler identification is GNU 10.1.0
The CXX compiler identification is GNU 10.1.0
Check for working C compiler: /usr/bin/cc
Check for working C compiler: /usr/bin/cc - works
Detecting C compiler ABI info
Detecting C compiler ABI info - done
Detecting C compile features
Detecting C compile features - done
Check for working CXX compiler: /usr/bin/c++
Check for working CXX compiler: /usr/bin/c++ - works
Detecting CXX compiler ABI info
Detecting CXX compiler ABI info - done
Detecting CXX compile features
Detecting CXX compile features - done
Building for Linux
Prepare lib and lib64 on Linux...
ln: failed to create symbolic link 'lib64': Permission denied
CMake Error at /usr/share/cmake-3.17/Modules/CMakeDetermineFortranCompiler.cmake:33 (message):
Could not find compiler set in environment variable FC:
CMAKE_Fortran_COMPILER-NOTFOUND.
Call Stack (most recent call first):
projects/openblas.cmake:3 (enable_language)
CMakeLists.txt:58 (include)
Configuring incomplete, errors occurred!
See also "/home/alsoussi/Projects/NewPrinter/cura/source/cura-build-environment/build/CMakeFiles/CMakeOutput.log".
and here is my CMakeOutput.log file
(The log file is too large so here is a github page)

How to use CMake

I am trying to rebuild this on my machine:
https://github.com/sfdodge/xnet
For this I needed to install xtensor and xtensor-blas. I followed the steps and successfully installed xtensor. However, I have trouble installing xtensor-blas. I create a fresh directory and run this command
cmake /home/fatima/Downloads/xtensor-blas-master
It generates the following log:
The C compiler identification is GNU 5.4.0
-- The CXX compiler identification is GNU 5.4.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- xtensor-blas v0.14.0
CMake Error at CMakeLists.txt:45 (find_package):
By not providing "Findxtensor.cmake" in CMAKE_MODULE_PATH this project has
asked CMake to find a package configuration file provided by "xtensor", but
CMake did not find one.
Could not find a package configuration file provided by "xtensor" with any
of the following names:
xtensorConfig.cmake
xtensor-config.cmake
Add the installation prefix of "xtensor" to CMAKE_PREFIX_PATH or set
"xtensor_DIR" to a directory containing one of the above files. If
"xtensor" provides a separate development package or SDK, be sure it has
been installed.
-- Configuring incomplete, errors occurred!
See also "/home/fatima/Downloads/Untitled Folder 3/CMakeFiles/CMakeOutput.log".
What am I doing wrong?

Finding GLib with CMake: target_include_directories called with invalid arguments

I have this following (minimal) CMakeLists.txt supposed to find GLib via pkg-config and add the libs to the foo target:
cmake_minimum_required(VERSION 2.10 FATAL_ERROR)
project(foo)
find_package(PkgConfig REQUIRED)
pkg_check_modules(GLib REQUIRED glib-2.0)
add_executable(foo foo.cpp)
message(WARNING "libs:" ${GLIB_LIBRARIES})
message(WARNING "includes:" ${GLIB_INCLUDE_DIRS})
target_link_libraries(foo PUBLIC ${GLIB_LIBRARIES})
target_include_directories(foo PUBLIC ${GLIB_INCLUDE_DIRS})
No matter what I try, I get (note the Found glib-2.0 part):
-- The C compiler identification is GNU 7.3.0
-- The CXX compiler identification is GNU 7.3.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found PkgConfig: /usr/bin/pkg-config (found version "0.29.1")
-- Checking for module 'glib-2.0'
-- Found glib-2.0, version 2.56.1
CMake Warning at CMakeLists.txt:6 (message):
libs:
CMake Warning at CMakeLists.txt:7 (message):
includes:
CMake Error at CMakeLists.txt:9 (target_include_directories):
target_include_directories called with invalid arguments
-- Configuring incomplete, errors occurred!
See also "/tmp/aaa/CMakeFiles/CMakeOutput.log".
and I cannot see, reading the CMake reference what arguments are invalid (note: this question is different from cmake target_include_directories called with invalid arguments). I also looked at CMake's FindPkgConfig documentation which gives glib as an example and I am not able to reproduce it (${GLIB_VERSION}). I tried GLIB_ and GLIB2_ prefixes and all I get is empty strings.
The messages show the variables are empty, though pkg-config reports values correctly:
$ pkg-config glib-2.0 --cflags --libs
-I/usr/include/glib-2.0 -I/usr/lib/x86_64-linux-gnu/glib-2.0/include -lglib-2.0
The CMake version I have is 2.10.
Can someone shed light on the issue?
I think your problem is the casing of GLIB_INCLUDE_DIRS and GLIB_LIBRARIES. They should be GLib_INCLUDE_DIRS and GLib_LIBRARIES since you specified "GLib" as the first argument of pkg_check_modules. I'm guessing that the target_include_directories() doesn't like not getting any arguments after the
PUBLIC (although it works for me in cmake 3.5).

CMake doesn't pass CUDA arguments?

I am trying to compile a library with both, C++ and CUDA source files. I am using GNU make with CMake. My compiler of choice is clang, since CUDA only supports gcc up to version 5 and Debian 9 has gcc 6 as its oldest version and I have to use software that is provided by the Debian 9 or 10 repositories.
CMake version is 3.9.0
clang version is 3.8.1
cc and c++ in /usr/bin correctly link to clang and clang++ ẃhich also link to the correct files.
Unfortunately the initial checks of CMake for CUDA fail although everything, as far as I can see, seems to be set up correctly. It looks like the arguments aren't passed correctly to the CUDA compiler.
This is a part of my project's main CMake file:
cmake_minimum_required (VERSION 3.9.0 FATAL_ERROR)
project (dev)
find_package(CUDA REQUIRED)
set(CUDA_HOST_COMPILATION_CPP ON)
set(CUDA_NVCC_FLAGS "${CUDA_NVCC_FLAGS} -Wno-deprecated-gpu-targets -ccbin clang-3.8")
The library's CMake file looks like this:
cmake_minimum_required(VERSION 3.9.0 FATAL_ERROR)
project (libname LANGUAGES CXX CUDA)
file(GLOB SOURCES "*.cu" "*.cpp")
add_library(libname ${SOURCES})
set_target_properties(libname PROPERTIES CUDA_SEPARABLE_COMPILATION ON)
set_target_properties(libname PROPERTIES POSITION_INDEPENDENT_CODE ON)
This is the output of CMake:
-- The C compiler identification is Clang 3.8.1
-- The CXX compiler identification is Clang 3.8.1
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Looking for pthread_create
-- Looking for pthread_create - not found
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE
-- Found CUDA: /usr (found version "8.0")
-- The CUDA compiler identification is unknown
-- Check for working CUDA compiler: /usr/bin/nvcc
-- Check for working CUDA compiler: /usr/bin/nvcc -- broken
The CUDA/nvcc tests of CMake fails with the following errors:
Change Dir: /home/user/projects/hamonIC-linux-experimental/current_state/working_copy/code/build/CMakeFiles/CMakeTmp
Run Build Command:"/usr/bin/make" "cmTC_6d1a9/fast"
/usr/bin/make -f CMakeFiles/cmTC_6d1a9.dir/build.make
CMakeFiles/cmTC_6d1a9.dir/build
make[1]: Entering directory
'/home/user/projects/hamonIC-linux-experimental/current_state/working_copy/code/build/CMakeFiles/CMakeTmp'
Building CUDA object CMakeFiles/cmTC_6d1a9.dir/main.cu.o
/usr/bin/nvcc -x cu -c
/home/user/projects/hamonIC-linux-experimental/current_state/working_copy/code/build/CMakeFiles/CMakeTmp/main.cu
-o CMakeFiles/cmTC_6d1a9.dir/main.cu.o
nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are
deprecated, and may be removed in a future release (Use
-Wno-deprecated-gpu-targets to suppress warning).
ERROR: No supported gcc/g++ host compiler found, but clang-3.8 is
available.
Use 'nvcc -ccbin clang-3.8' to use that instead.
CMakeFiles/cmTC_6d1a9.dir/build.make:65: recipe for target
'CMakeFiles/cmTC_6d1a9.dir/main.cu.o' failed
make[1]: *** [CMakeFiles/cmTC_6d1a9.dir/main.cu.o] Error 1
make[1]: Leaving directory
'/home/user/projects/hamonIC-linux-experimental/current_state/working_copy/code/build/CMakeFiles/CMakeTmp'
Makefile:126: recipe for target 'cmTC_6d1a9/fast' failed
make: *** [cmTC_6d1a9/fast] Error 2
CMake will not be able to correctly generate this project.
Call Stack (most recent call first):
hic_iplibrary/source/ciccommon/CMakeLists.txt:7 (project)
The CUDA NVCC flags in CMake have to be semicolon delimited and not whitespace delimited.
Change the flags in your CMakeLists.txt to use:
set(CUDA_NVCC_FLAGS "${CUDA_NVCC_FLAGS};-Wno-deprecated-gpu-targets;-ccbin=clang-3.8")
and that should allow CMake to pass your flags to NVCC.