I am trying to capitalize on the sample applications provided for Nano.
CUDA samples: I was able to use Nsight eclipse compiler to build for x86_64 and run on aarch64(Nano). This also allowed me to debug the application through SSH(Cross-compile).
VisionWorks/OpenVX API samples: Now, in order to use PiCam with GSTREAMER->EGL->Tegra to capture live stream and process them, the demos/samples are a part of VisionWorks/OpenVX API samples. While i am able to build and run these demos on Nano itself in nativemode, i fail to import them on Nsight Eclipse other than manual import as makefile projects, but the build requires Tegra RootFS for x86. The debian packages i received as CUDA toolkit/host components in SDK manager only has Tegra Rootfs for aarch64.
Is there any directive to import and build/debug over ssh for the visionworks samples just like CUDA samples in cross-compile format as none of our organization laptops have GPUs. How can i debug/remote debug visionworks samples/applications while runtime on Nano.
Is it possible to build an application in Nsight Eclipse that uses both CUDA apis(for processing) + VisionWorks/GStreamer/NVXIO api(for HW interactions) to access CSI cameras or other Tegra HW components.
Note: I had to tag this topic under CUDA/GSTREAMER as there was no VisionWorks/OpenVX Tag available.
Alike CUDA samples, can Nsight Eclipse can be used to Build/Debug/Profile VisionWorks Samples on x86 host in Cross-compile setup?