hi,
I would like to cross-compile on Ubuntu20.04 to my Jetson board,Can cuda cross-compile with arm64-sbsa? Or which compiler should I download if I want to implement cuda cross-compile?
thanks
Hi,
arm64-sbsa is a different system.
Please install the cross-compile toolkit from the SDK manager.
The package is also available in the below link: cuda-cross-aarch64*
https://repo.download.nvidia.com/jetson/
Thanks.
If I install CUDA-cross-aARCH64 successfully, can I cross-compile cuda?
If cuda-cross-aarch64 is installed successfully, where can I see a successful installation sign?
Hi,
You can check it with $ aarch64-linux-gnu-gcc -v
.
Thanks.
Do I cross-compile cuda using nvcc or aarch64-linux-gnu-gcc or CUDA-Cross-SBSA?
hi,Can I find nvcc when I install cuda cross compiler without cuda
Hi,
You will also need a CUDA toolkit.
GPU code should be complied with nvcc and the corresponding flag.
Ex. -gencode arch=compute_72,code=sm_72
for Xavier.
CPU code should be complied with aarch64-linux-gnu-gcc.
Thanks.
hi,
How should I configure cmakelist when compiling a project with cuda code?
thanks.
hi,
Are there any cuda cross-compilation examples you can share with me? I’d like to see how your cmakelist is configured
thanks.
nvcc fatal : Unknown option '-gencode arch=compute_72,code=sm_72
After I add this sentence, I will prompt this error, do you know how to solve this error, or what caused this error?
Hi,
Please find an example below:
https://docs.nvidia.com/vpi/sample_cross_aarch64.html
Add set the GPU architecture with the following command:
set(CUDA_NVCC_FLAGS ${CUDA_NVCC_FLAGS}; -O3)
set(CUDA_NVCC_FLAGS ${CUDA_NVCC_FLAGS}; -gencode arch=compute_72,code=sm_72)
Thanks.
error adding symbols: file in wrong format
Do you know how to solve this error when cross-compiling cuda programs
There is no update from you for a period, assuming this is not an issue any more.
Hence we are closing this topic. If need further support, please open a new one.
Thanks
Hi,
Do you build it in a clean environment?
Could you run make clean
and try it again?
Thanks.