I’m trying to profile a program with nsight-compute on host Windows machine(without CUDA devices) and connect to a linux device. However, the program can not be launched with the error note:
Could not find required target platform directory C:/program files/NVIDIA Corporation/Nsight Compute 2019.1/target/linux-desktop-glibc_2_11_3-x64
I don’t know if there is some issue with installation, but I don’t see any options about that.
Thanks!
kirlia
Remote launching from Windows to Linux is not yet fully supported. It will be added in a future release. In the meantime, there are two ways to address this.
You can install the Linux version of Nsight Compute on the Linux target machine, launch the application using nv-nsight-cu-cli there, and attach from the Windows host.
https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html#modes
You can install the Linux version of Nsight Compute on a Linux machine, and copy the target/linux-desktop-glibc_2_11_3-x64 directory to your Windows host installation.