Hi,
I was looking at this talk from one of backend developers of Microsoft's c++ compiler. link: https://youtu.be/Edzh_e1rqCo?t=40m49s

He was talking about parallelizing control flow and why this optimization cant be done to c++ code. Spoiler: because c++ variables are shared across threads. Since D is selfish and doesn't share by default 99% of cases could be optimized by D compiler.

Just wanted to make sure that some one knew about it and such optimization would be implemented in the future. Bye