This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH RFC: Speed up compilation of 20001226-1.c
Jan Hubicka wrote:
> If engline was keeping track of changed edges and calling VRP per edge
> basis, we would have linear time complexity as promised. I've started
> implementing this instead of the trick with two queues but got
> sidetacked by IPA merge and I don't think I have the patch around
> anymore, so I guess the simple heuristics might be good pick.
Based on this discussion, I think Ian's patch is OK.
Everyone's agreed it's correct, and nobody thinks it's harmful. The
20001226-1.c test accounts for a noticeable fraction of overall test
time, so speeding it up is a win, if just for us GCC developers. There
may be better improvements, but none of them are ready, and Ian's change
doesn't make it any harder to implement them.
So, Ian, I think you should check in your patch.
Thanks,
--
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713