windows support?

windows support?

Thanks. NVIDIA was quite generous in open-sourcing this as it is.

@sscalpone If you guys set up an appveyor account on this repository, I might be able to set up automated builds for windows. Since building the entire stack will take more than the one hour limit, the idea is to cache the build files and then build the next thing if you have time.

Libraries:

• llvm from vcpkg
• openmp-llvm from conda
• clang-flang needs to be compiled manually

Then of course the flang build would come from this repository.

I'm going to try to get this working on my appveyor account first, and if it works, then I can transfer it here. Currently no ETA on completion, as I may run into issues that will require more time than I currently have.

[sscalpone]

You'll see that the code has #ifdefs for Windows. The bulk of the initial work will be with CMakefiles, and then filling in the run-time libraries. NVIDIA doesn't have near-term plans to do this work.

Update: this comment was completely wrong, sorry.