Hello, there are dumps for Inkscape, it looks very well. Link: https://docs.google.com/file/d/0B0pisUJ80pO1Y0t1aEVBRlByR28/edit
There are few of functions that look like this (wpa cgraph): _ZL13resync_activeP19_EgeSelectOneActionii/2604322 (resync_active) @0x7f84af42cea0 Type: function definition analyzed Visibility: prevailing_def_ironly References: Referring: Read from file: libinkscape.a Availability: local First run: 4422 Function flags: executed 47x local Called by: ege_select_one_action_set_active_text/2604300 (0.34 per call) (can throw external) _ZL21commit_pending_changeP19_EgeSelectOneAction/2604327 (0.16 per call) (can throw external) _ZL34ege_select_one_action_set_propertyP8_GObjectjPK7_GValueP11_GParamSpec/2604316 (47x) (0.16 per call) (can throw external) Calls: _ZL13resync_activeP19_EgeSelectOneActionii.part.0/2604456 (10x) (0.21 per call) (can throw external) _ZL13resync_activeP19_EgeSelectOneActionii.part.0/2604456 (_ZL13resync_activeP19_EgeSelectOneActionii.part.0) @0x7f84af42cd68 Type: function definition analyzed Visibility: artificial References: _ZL7signals/2604291 (read) Referring: Read from file: libinkscape.a Availability: local First run: 0 Function flags: executed 10x local Called by: _ZL13resync_activeP19_EgeSelectOneActionii/2604322 (10x) (0.21 per call) (can throw external) First function has a profile (position is correct according to valgrind) and second not. Both of them comes from the same object file. The problem is that the second one is called according to valgrind. What does .part.X means, is it a part of function that was separated to a different function? Is there any was these two profiles could be merged? Thank you, Martin On 27 November 2013 00:24, Martin Liška <marxin.li...@gmail.com> wrote: > Hello, > present results reached for GIMP by the reordering pass. Important > to notice, that I just used single '.text' section where all symbols > are placed. As you can see, there just few functions that are not > catched by the pass (3 of them are LTO clones, that I will find out). > And 2 functions were not seen during -fprofile-generate run. > > In following days, I will prepare same dumps for Inkscape and Firefox. > > Martin > > On 18 November 2013 11:16, Jan Hubicka <hubi...@ucw.cz> wrote: >>> diff --git a/gcc/ChangeLog b/gcc/ChangeLog >>> index 5cb07b7..754f882 100644 >>> --- a/gcc/ChangeLog >>> +++ b/gcc/ChangeLog >>> @@ -1,3 +1,13 @@ >>> +2013-11-17 Martin Liska <marxin.li...@gmail.com> >>> + Jan Hubicka <j...@suse.cz> >>> + >>> + * cgraphunit.c (node_cmp): New function. >>> + (expand_all_functions): Function ordering added. >>> + * common.opt: New profile based function reordering flag introduced. >>> + * lto-partition.c: Support for time profile added. >>> + * lto.c: Likewise. >>> + * predict.c (handle_missing_profiles): Time profile handled in >>> + missing profiles. >> >> OK, >> thanks! Implementing the function section naming scheme would be easy and >> it would >> enable us to do the reordering even w/o LTO that would be quite cool. Lets >> hope it gets >> resolved soon. >> >> Honza
