Sorry, but the Windows HPC SDK is still in development so isn’t available. We hope to release it sometime later this year.
In the meantime, we have had customers install the Linux package under Windows WSL2 (Windows Subsystem for Linux). We don’t officially support WSL2 yet since we haven’t thoroughly tested it, but it might be worth a try.
FYI for your StackOverflow posts, wish I could help more but I’m unclear if GNU supports OpenACC in Cygwin or MinGW. But your code, sans the data region which is incorrect, works fine for me with nvc++ on Linux. I just use the flag “-gpu=managed” to enable CUDA Unified Memory so that the CUDA driver manages the data movement. I can respond over there if you like, but was hoping some else who would have more knowledge of GNU would jump in.