Remote application development using NVIDIA® Nsight™ Eclipse Edition

Hi kevin kang,
Thank you,l can find the "hammer",which means "chuizi" in Chinese.
But l come across a problem:
After l click debug under the jetson tk1(my target system),there is a error,the detail is:

Errors occurred during the build.
Errors running builder 'CDT Builder' on project 'particle'.
Internal error building project particle configuration Debug
org.eclipse.ptp.rdt.sync.core.exceptions.RemoteSyncException: Remote Git commit failed with message:
Remote Git commit failed with message:
Internal error building project particle configuration Debug
org.eclipse.ptp.rdt.sync.core.exceptions.RemoteSyncException: Remote Git commit failed with message:

what should l do?

Hi everyone,
After l click on the build “hammer” icon in the toolbar to drop down the build menu.Then,
l come across a problem:

Errors occurred during the build.
Errors running builder 'CDT Builder' on project 'particle'.
Internal error building project particle configuration Debug
org.eclipse.ptp.rdt.sync.core.exceptions.RemoteSyncException: Remote Git commit failed with message:
Remote Git commit failed with message:
Internal error building project particle configuration Debug
org.eclipse.ptp.rdt.sync.core.exceptions.RemoteSyncException: Remote Git commit failed with message:

what should l do?

Seems like you're hitting an Eclipse known issue below:
https://bugs.eclipse.org/bu...

The similar issue(along with solution) was posted on Devtalk form as following:
https://devtalk.nvidia.com/...

Hope it helps.

Sync'ing should be done by using a separate Git directory (.ptp-sync).

Please refer to:
https://devtalk.nvidia.com/...
And,
https://bugs.eclipse.org/bu...

Hello Satish,

I am working on Jetson TX1 kit. I installed ubuntu 16.06 with gcc5. It has an 64-bit ARM® A57 CPUs. I am trying to use nvidia's nsight for cross-compilation. I am doing development work related to point gray camera on ARM. I am using " remote synchronized project mode" way of cross compilation using cuda tool kit. I will be using ARM build libraries along with cuda built libraries. While cross-compiling, it shoots "undefined references : VTT for std::__cxx11::basic_stringstream<char, std::char_traits<char="">, std::allocator<char> >@GLIBCXX_3.4.21'" and many such undefined references errors.

I am attaching the snapshots of the setting I did for the program.

It would be highly appreciated if you throw some light on my endeavour

https://uploads.disquscdn.c... https://uploads.disquscdn.c... https://uploads.disquscdn.c...

From Nsight EE, the linker library search path(-L) and the linker libraries(-l) can be set using the project property page.

Right click on the project-> Properties->Build->Settings->NVCC Linker->Libraries.

Also C++11 support can be enabled using the following option,

Right click on the project-> Properties->Build->Settings->NVCC Compiler->Code Generation->Enable C++11 support.

I' not sure if the above will help you with your compiler errors -- without seeing code I'm not sure what would be causing them.

Hi, I'm recently building on remote Tx1 board and facing the same issue "Permission denied". Is that issue fixed? Thanks

Dear Satish,
Thank you so much for this post. This is really useful in using Jetson TX1 boards since it dosn't have alot of memory to carry IDE like Nsight....However I am wondering if this will work with PyCuda code which requires a python IDE, and I tried to install Pydev for Nsight which dosn't work so I am just curious if there is any other alternatives to remote compute without having Nsight?? THank you so much

Hi Daniel, Can you post the exact error message from the console view? Are you seeing the permission denied error when trying to run/debug the application on target? Please also check if the target directory has the appropriate permission to copy the executable.

Last login: Thu Jan 26 04:31:14 2017 from 192.168.2.14
echo $PWD'>'
/bin/sh -c "cd \"/home/ubuntu/Documents\";export LD_LIBRARY_PATH=\"/usr/local/cuda-8.0/lib64\":\${LD_LIBRARY_PATH};\"/home/ubuntu/Documents/Remote_App\"";exit
ubuntu@tegra-ubuntu:~$ echo $PWD'>'
/home/ubuntu>
ubuntu@tegra-ubuntu:~$ /bin/sh -c "cd \"/home/ubuntu/Documents\";export LD_LIBRARY_PATH=\"/usr/local/cuda-8.0/lib64\":\${LD_LIBRARY_PATH};\"/home/ubuntu/Documents/Remote_App\"";exit
/bin/sh: 1: /home/ubuntu/Documents/Remote_App: Permission denied
logout

Above is error message. Btw, I use Wireless connection from host to tx1

Hi,I have jetson tx1 ,I got the following error. How can I solve it?

Last login: Mon Mar 13 20:53:11 2017 from 192.168.1.34
echo $PWD'>'
/bin/sh -c "cd \"/home/ubuntu/Desktop/Denemeler/\";export LD_LIBRARY_PATH=\"/usr/local/cuda-8.0/lib64\":\${LD_LIBRARY_PATH};\"/home/ubuntu/Desktop/Denemeler/remoteprojects/\"";exit
ubuntu@tegra-ubuntu:~$ echo $PWD'>'
/home/ubuntu>
ubuntu@tegra-ubuntu:~$ /bin/sh -c "cd \"/home/ubuntu/Desktop/Denemeler/\";export LD_LIBRARY_PATH=\"/usr/local/cuda-8.0/lib64\":\${LD_LIBRARY_PATH};\"/home/ubuntu/Desktop/Denemeler/remoteprojects/\"";exit
/bin/sh: 1: /home/ubuntu/Desktop/Denemeler/remoteprojects/: Permission denied
logout

Hi, I am unsure of the error, can you please post to Jetson TX1 Technical Support Forum? https://devtalk.nvidia.com/...
Thanks

set Git before following steps

Thanks for the tutorial. On my nsight release 7.5 running on MacOS X, I get include file errors for all include files (e.g. stdio.h) in the editor. The code compiles and runs fine on the target, which is Linux CentOS. The project explorer does not show any include files on the remote system, as it usually does for the local system. The include errors are also falsely flagging many other data structures. Is there way to stop the include errors in the editor and have the project include the correct files on the remote system?
Thanks.

https://uploads.disquscdn.c...

Can you try running the indexer again to see if that resolves the include error?
Right click on the project-> Index-> Rebuild

You can specify the include paths in project property page (Build->Settings->Tool settings..->NVCC compiler->includes)

I have the same problem. But the eclipse doesn't support indexing include files on a remote system. My solution is to install the same version of gcc and other dependencies on local system, and set PATH and CPATH environment variable. But recently I found that maybe RDT can be used to index remote include files.
https://uploads.disquscdn.c...

How to run CUDA sample programs using Nsight Eclipse Edition in Ubuntu OS. Please explain in detail. Every time I m getting error in debugging. There is no more helpful lecture regarding how to run cuda samples program using Nsight.

Hi,

I was able to cross compile the inbuilt CUDA Samples and also my own helloworld code.

However, I am working on a SLAM algorithm where I have lots of src file and Thirdparty software files that needs to be linked. I am not able to link a big project where I want only single executable but I have lots of source and include files.
tx1.cpp should be the executable.
This is the error.

Finished building: /home/apoorva/ORB-SLAM2-GPU2016-final/gpu/tx1.cpp

Building target: libtest_cuda_SLAM
Invoking: NVCC Linker
/usr/local/cuda-8.0/bin/nvcc --cudart shared -L"/usr/local/cuda-8.0/samples/common/inc" -L"/usr/local/cuda-8.0/samples/common/lib/linux/aarch64" -L"/usr/local/cuda-8.0/targets/aarch64-linux/lib" -L"/home/apoorva/ORB-SLAM2-GPU2016-final/cmake_modules" -L"/usr/local/lib" -shared -lGL -lGLU -lglut -Xlinker -unresolved-symbols=ignore-in-shared-libs --relocatable-device-code=true -gencode arch=compute_60,code=compute_60 -gencode arch=compute_60,code=sm_60 -m64 -ccbin aarch64-linux-gnu-g++ -link -o "libtest_cuda_SLAM" ./src/Allocator_gpu.o ./src/Converter.o ./src/Cuda.o ./src/Fast_gpu.o ./src/Frame.o ./src/FrameDrawer.o ./src/Initializer.o ./src/KeyFrame.o ./src/KeyFrameDatabase.o ./src/LocalMapping.o ./src/LoopClosing.o ./src/Map.o ./src/MapDrawer.o ./src/MapPoint.o ./src/ORBextractor.o ./src/ORBmatcher.o ./src/Optimizer.o ./src/Orb_gpu.o ./src/PnPsolver.o ./src/Sim3Solver.o ./src/System.o ./src/Tracking.o ./src/Viewer.o ./gpu/tx1.o /usr/local/cuda-8.0/samples/common/lib/linux/x86_64/libGLEW.a -lcudart -lopencv_core -ltbb -lz
/usr/lib/gcc-cross/aarch64-linux-gnu/5/../../../../aarch64-linux-gnu/bin/ld: skipping incompatible /usr/local/lib/libopencv_core.so when searching for -lopencv_core
makefile:60: recipe for target 'libtest_cuda_SLAM' failed
/usr/lib/gcc-cross/aarch64-linux-gnu/5/../../../../aarch64-linux-gnu/bin/ld: skipping incompatible //usr/local/lib/libopencv_core.so when searching for -lopencv_core
/usr/lib/gcc-cross/aarch64-linux-gnu/5/../../../../aarch64-linux-gnu/bin/ld: cannot find -lopencv_core
collect2: error: ld returned 1 exit status
make: *** [libtest_cuda_SLAM] Error 1

Hi Satish, I am trying to use Nsight to create a synchronized remote project but as I try to build I get the following error (see attached Figure):

https://uploads.disquscdn.c...

Any idea? My GIT version on remote is the 2.5 while on local is 2.7.4. Can this be an issue? Many thanks

Jony

Hi Jony,

It looks like the git is not configured properly(the error says git command not found). Have you used git config command as specified above? Also more information about this error can be found in the log file($your_workspace/.metadata/.log). Can you send the log related to RemoteExecutionException?