Possibility of building CUDA/openCL results from x86 machine running on Jetson Nano?

Hello all,

Assume that someone build .cu or .cl ( CUDA / openCL ) on a machine with x86 CPU and 30 series GPU
and generate the binary file with cross-compile-toolchain such as LLVM-Clang.

Does the binary file also work correctly on Jetson series platform such as Nano and Xavier NX ?

Thanks for your reply !!

Hi,

You will need to use cross-compile to make sure the binary can be executed on the ARM system.
And for Nano, please also add GPU architecture 53 to the nvcc.

https://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html#options-for-steering-gpu-code-generation-gpu-architecture

Thanks.