On 10 September 2013 11:10, deadalnix <deadal...@gmail.com> wrote:
> On Monday, 9 September 2013 at 16:43:54 UTC, Andrei Alexandrescu wrote:
> So I'll jump in as I've seen the conf before that post.
> It is really worthwhile to watch. Quite informative on the technical side
> (funilly, I did implement the solution presented in the devirtualization
> part on a platform where virtual function were not available because of
> crappy compiler, but had no idea it was worthwhile for speed, it wasn't the
> goal so I never measured, and it also may not the case on the given
Interestingly enough, gcc recently added (about a month before Andrei
did the talk) a new interprocedural analysis pass that sets all
methods that can be devirtualised. I expect that LLVM have something
cooking up for this too. As of yet, I haven't ran any tests which
show that this works though...
*(p < e ? p++ : p) = (c & 0x0f) + '0';