Having below linker errors, not entirely sure what/which paths need to be added?
g++ -O2 -std=c++17 -I/usr/local/cuda-11.4/include -I/home/^^^/Desktop/jetson-inference/utils -I/home/^^^/Desktop/jetson-inference/utils/cuda goai.cpp -c
g++ -L/usr/local/cuda/lib64 -L/home/^^^/Desktop/jetson-inference/build/aarch64/lib -lcudart -ljetson-inference -ljetson-utils goai.o -o TestExecs
errors
Hi,
Are you building jetson-utils and jetson-inference on the Nano directly?
The DSO missing error usually indicates that the file is compiled with a difference architecture (ex. ARM vs. AMD).
Thanks.
Yes on the Nano (Jetson nano orin)
Building it from source, following the steps provided herehttps://github.com/dusty-nv/jetson-inference/blob/master/docs/building-repo-2.md ]
in /usr/local/lib the two static objects are there, post sudo make install
i guess it was sudo ldconfig (which probably has went wrong?)
/usr/local/lib:
/sbin/ldconfig.real: Can’t create temporary cache file /etc/ld.so.cache~: Permission denied
Hi @cedricyangks , not sure how you have CUDA 11.4 on JetPack 4 and if that will work or not, but here is an example CMakeLists project for linking to jetson-inference/jetson-utils externally:
# require CMake 2.8 or greater
cmake_minimum_required(VERSION 2.8)
# declare my-recognition project
project(my-recognition)
# import jetson-inference and jetson-utils packages.
# note that if you didn't do "sudo make install"
# while building jetson-inference, this will error.
find_package(jetson-utils)
find_package(jetson-inference)
# find VPI package (optional)
find_package(VPI 2.0)
# CUDA is required
find_package(CUDA)
# add directory for libnvbuf-utils to program
show original
the CMake and Makefile approach works (followed the guide on github), but on my own exploration i can’t get the linker to work.
By analogy, if i can link it with CMake if i set my paths correct in the linker it should work as well?
**g++ -O2 -std=c++17 -I/usr/local/cuda/include -I/home/xxx/Desktop/jetson-inference/build/aarch64/include -I/home/xxx/Desktop/jetson-inference/utils/cuda goai.cpp -c**
**g++ -L/usr/local/lib -L/usr/local/cuda/lib64 -L/home/xxx/Desktop/jetson-inference/build/aarch64/lib -lcudart -ljetson-inference -ljetson-utils goai.o -o TestExecs**
Using jetpack version 5.1.2
:~/Desktop$ apt show nvidia-jetpack
Package: nvidia-jetpack
Version: 5.1.2-b104
If I compile the my-recognition example on JetPack 5.1.2 with that CMakeLists, this is the command-line that gets run with VERBOSE=1 make
/usr/bin/c++ -I/usr/local/cuda-11.4/include -isystem /opt/nvidia/vpi2/include -o CMakeFiles/my-recognition.dir/my-recognition.cpp.o -c /mnt/NVME/jetson-inference/examples/my-recognition/my-recognition.cpp
[100%] Linking CXX executable my-recognition
/usr/bin/cmake -E cmake_link_script CMakeFiles/my-recognition.dir/link.txt --verbose=1
/usr/bin/c++ -rdynamic CMakeFiles/my-recognition.dir/my-recognition.cpp.o -o my-recognition -L/usr/lib/aarch64-linux-gnu/tegra -Wl,-rpath,/usr/lib/aarch64-linux-gnu/tegra:/usr/local/lib:/usr/local/cuda-11.4/lib64:/opt/nvidia/vpi2/lib/aarch64-linux-gnu /usr/local/cuda-11.4/lib64/libcudart_static.a -lpthread -ldl -lrt /usr/local/lib/libjetson-inference.so /usr/local/lib/libjetson-utils.so /usr/local/cuda-11.4/lib64/libcudart_static.a -lpthread -ldl -lrt -lGL -lGLU -lGLEW -lgstreamer-1.0 -lgstapp-1.0 -lgstpbutils-1.0 -lgstwebrtc-1.0 -lgstsdp-1.0 -lgstrtspserver-1.0 -ljson-glib-1.0 -lsoup-2.4 /usr/local/cuda-11.4/lib64/libnppicc.so -lnvinfer -lnvinfer_plugin -lnvcaffe_parser -lnvonnxparser -lopencv_core -lopencv_calib3d /opt/nvidia/vpi2/lib/aarch64-linux-gnu/libnvvpi.so.2.3.8
Since jetson-inference is installed under /usr/local , I don’t believe you should have to specify additional paths. As in the my-recognition sample though, headers do need to be included like:
#include <jetson-inference/imageNet.h>
#include <jetson-utils/loadImage.h>
tyvm - managed to have it resolved.
g++ -O2 -std=c++17 -I/usr/local/cuda/include -I/home/x3d/Desktop/jetson-inference/build/aarch64/include -I/home/x3d/Desktop/jetson-inference/utils/cuda goai.cpp -c
g++ goai.o -o my-imagnetExe3 -L/usr/lib/aarch64-linux-gnu/tegra -L/usr/local/lib -L/usr/local/cuda-11.4/lib64 -lpthread -ldl -lrt -ljetson-inference -ljetson-utils