On Tue, Dec 11, 2012 at 1:43 AM, Richard Biener <richard.guent...@gmail.com> wrote: > On Mon, Dec 10, 2012 at 8:55 PM, Sharad Singhai <sing...@google.com> wrote: >> On Mon, Dec 10, 2012 at 11:02 AM, Richard Biener >> <richard.guent...@gmail.com> wrote: >>> On Mon, Dec 10, 2012 at 7:46 PM, Sharad Singhai <sing...@google.com> wrote: >>>> Hi, >>>> >>>> The new dump infrastructure was committed shortly before the trunk >>>> entered stage 3. >>>> >>>> However, except the vectorization passes, other passes do not dump >>>> anything in response to -fopt-info flags despite the documentation. I >>>> can prepare patches for a couple more passes so that they output more >>>> meaningful information, otherwise, the -fopt-info would remain largely >>>> useless in 4.8. Also I suspect some other passes can be converted >>>> relatively easily to use the new dump scheme. >>>> >>>> Would it be acceptable to consider -fopt-info patches in stage 3? If >>>> yes, I can send out a more detailed message about the conversion and >>>> solicit for help. >>> >>> It should be decided on a case-by-case basis. What passes do you have in >>> mind >>> that people are usually interested in in the context of -fopt-info? >> >> For starters, I was looking at unrolling and inlining decisions >> (loop-unroll.c and tree-inline.c respectively). > > Unrolling is probably easy, I'm not sure how to handle inlining (the info > will be very noisy for C++ for example, so eventually early inlining should > be excempt). Make sure to propose separate patches for different passes > you handle - giving examples of -fopt-info output for them. I do think that > exposing more stuff via it for 4.8 would be nice.
Yes, unrolling to use new dump scheme was quite easy. I proposed a patch in http://gcc.gnu.org/ml/gcc-patches/2012-12/msg00953.html. I think I will hold off on inlining as it seems more involved. I looked at if-conversion and that looks like a ripe candidate for next dump conversion. Thanks, Sharad
