Anastasia accepted this revision. Anastasia added a comment. Your code comment seem to describe the issue quite well. Even though I am still inclined towards keeping the address spaces as long as possible and only converting into physical memory segments on the backend really. I believe there are a couple of flatten memory architectures that do that and it seems to work fine. But on the other hand it seems that there are other implementations that took the path of flattening directly in Clang and therefore are facing the issue you are solving now.
I will try to send some draft on OpenCL documentation in Clang hopefully the following weeks. Perhaps it will help us to converge to the common flow at some point. Repository: rL LLVM https://reviews.llvm.org/D26157 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits