I have a Linux (x64) version of cuda-6.0 with nvcc that claims to support --target-cpu-architecture ARM. On the other hand, the latest Windows nvcc does not recognize ARM. Is this my problem, like I did not install the necessary packages, or does NVIDIA only provide such cross-compiler for Linux hosts?
The cross-compile targets for any given OS are listed in the relevant getting started documents. Refer to tables 1 and 2 in the Windows and Linux documents respectively.
You cannot target ARM from a windows-based CUDA toolkit.
Collected from the Internet
Please contact [email protected] to delete if infringement.
Comments