Errors compiling ndt_omp

[amlobat2]

I am new to ROS and Linux and can't seem to figure out why I am getting issues when using catkin_make to compile the ndt_omp requirement. The error I get is below.

In file included from /media/oliver/amlobat2_files/hdl_graph_slam/src/ndt_omp/src/pclomp/gicp_omp.cpp:2:0:
/media/oliver/amlobat2_files/hdl_graph_slam/src/ndt_omp/include/pclomp/gicp_omp_impl.hpp: In member function ‘void pclomp::GeneralizedIterativeClosestPoint<PointSource, PointTarget>::computeCovariances(typename pcl::PointCloud<PointTDiff>::ConstPtr, typename pcl::search::KdTree<PointT>::ConstPtr, pclomp::GeneralizedIterativeClosestPoint<PointSource, PointTarget>::MatricesVector&)’:
/media/oliver/amlobat2_files/hdl_graph_slam/src/ndt_omp/include/pclomp/gicp_omp_impl.hpp:64:51: error: there are no arguments to ‘omp_get_max_threads’ that depend on a template parameter, so a declaration of ‘omp_get_max_threads’ must be available [-fpermissive]
   std::vector<std::vector<int>> nn_indecies_array(omp_get_max_threads());
                                                  ^~~~~~~~~~~~~~~~~~~
/media/oliver/amlobat2_files/hdl_graph_slam/src/ndt_omp/include/pclomp/gicp_omp_impl.hpp:64:51: note: (if you use ‘-fpermissive’, G++ will accept your code, but allowing the use of an undeclared name is deprecated)
/media/oliver/amlobat2_files/hdl_graph_slam/src/ndt_omp/include/pclomp/gicp_omp_impl.hpp:65:52: error: there are no arguments to ‘omp_get_max_threads’ that depend on a template parameter, so a declaration of ‘omp_get_max_threads’ must be available [-fpermissive]
   std::vector<std::vector<float>> nn_dist_sq_array(omp_get_max_threads());
                                                      ^~~~~~~~~~~~~~~~~~~
media/oliver/amlobat2_files/hdl_graph_slam/src/ndt_omp/include/pclomp/gicp_omp_impl.hpp:69:43: error: there are no arguments to ‘omp_get_thread_num’ that depend on a template parameter, so a declaration of ‘omp_get_thread_num’ must be available [-fpermissive]
     auto& nn_indecies = nn_indecies_array[omp_get_thread_num()];
                                           ^~~~~~~~~~~~~~~~~~
/media/oliver/amlobat2_files/hdl_graph_slam/src/ndt_omp/include/pclomp/gicp_omp_impl.hpp:70:41: error: there are no arguments to ‘omp_get_thread_num’ that depend on a template parameter, so a declaration of ‘omp_get_thread_num’ must be available [-fpermissive]
     auto& nn_dist_sq = nn_dist_sq_array[omp_get_thread_num()];
                                         ^~~~~~~~~~~~~~~~~~

I get this same error for each instance that omp_get_max_threads and omp_get_thread_num. Can someone please help me figure out what to do? Thanks!!

[fghjkl17]

I want to know which architecture does your CPU belong to? amd64 or arm64 or else?

[amlobat2]

I'm working on an NVIDIA Xavier so it's armv8

[fghjkl17]

Could you change line 5-6 in hdl_graph_slam/CMakeLists.txt and ndt_omp/CMakeLists.txt from add_definitions(-std=c++11 -msse -msse2 -msse3 -msse4 -msse4.1 -msse4.2) set(CMAKE_CXX_FLAGS "-std=c++11 -msse -msse2 -msse3 -msse4 -msse4.1 -msse4.2") to add_definitions(-std=c++11) set(CMAKE_CXX_FLAGS "-std=c++11") and then delete build and devel folders and catkin_make again.

[ipinyulun]

error_Tx2.txt

Could you change line 5-6 in hdl_graph_slam/CMakeLists.txt and ndt_omp/CMakeLists.txt from add_definitions(-std=c++11 -msse -msse2 -msse3 -msse4 -msse4.1 -msse4.2) set(CMAKE_CXX_FLAGS "-std=c++11 -msse -msse2 -msse3 -msse4 -msse4.1 -msse4.2") to add_definitions(-std=c++11) set(CMAKE_CXX_FLAGS "-std=c++11") and then delete build and devel folders and catkin_make again.

i encunter the same issue (as file uploaded up there) and i try this solution, however it does not work. one of my concern is PCL1.8. becoz im running on ubuntu 18.04 on Jetson Tx2. aarch64

so i was unable to install PCL 1.7

therefore, is the cause of the error becoz of PCL1.8 ? if yes, how to install PCL 1.7 into Jetson Tx2? google doesnt seem help as well

[ipinyulun]

error_Tx2.txt

Could you change line 5-6 in hdl_graph_slam/CMakeLists.txt and ndt_omp/CMakeLists.txt from add_definitions(-std=c++11 -msse -msse2 -msse3 -msse4 -msse4.1 -msse4.2) set(CMAKE_CXX_FLAGS "-std=c++11 -msse -msse2 -msse3 -msse4 -msse4.1 -msse4.2") to add_definitions(-std=c++11) set(CMAKE_CXX_FLAGS "-std=c++11") and then delete build and devel folders and catkin_make again.

i encunter the same issue (as file uploaded up there) and i try this solution, however it does not work. one of my concern is PCL1.8. becoz im running on ubuntu 18.04 on Jetson Tx2. aarch64

so i was unable to install PCL 1.7

therefore, is the cause of the error becoz of PCL1.8 ? if yes, how to install PCL 1.7 into Jetson Tx2? google doesnt seem help as well

issue solved at https://github.com/koide3/ndt_omp/issues/2

[FishInWave]

@ipinyulun Hi, thanks for your quote. I successfully run hdl_graph_slam on my Jetson Xavier with a rosbag recorded by a robosense-16 LiDar. But I note that the SLAM cannot run in 1X speed normally even when Xavier is in "MODE 30W ALL"，i.e. 8-core is activeted, and I must use record paly -r 0.5 to slow down play speed. I wonder whether you can run this program on your TX2 in real time?