Trying to use nsight to select remote agx target but getting error

I’m remoting in from a linux x86.

I see this error on the NVIDIA nsight GUI:

    OriginalExceptionClass: N5boost16exception_detail10clone_implIN11QuadDCommon26DirectoryNotFoundExceptionEEE
    OriginalFile: /build/agent/work/20a3cfcd1c25021d/QuadD/Host/Analysis/PosixDeviceValidator.cpp
    OriginalLine: 24
    OriginalFunction: bool QuadDAnalysis::PosixDeviceValidator::CheckHostSupport(const DevicePtr&)
    Filename: /opt/cuda/nsight_systems/target-linux-tegra-armv8
    ErrorText: Deploy directory does not exist

Seems like I need to install arm stuff on my x86 system to remote into the agx. How would I do that if that 's the actual issue?


Do you use the Nsight tool from the same JetPack installer?

If not, please give it a try.
It also includes the cross-compiling toolkit that is required for the remote profiling.