[Bug tree-optimization/18419] New: [4.0 Regression] tree-ssa aliasing slow

[Bug tree-optimization/18419] New: [4.0 Regression] tree-ssa aliasing slow

[pinskia at gcc dot gnu dot org]

The testcase from PR 16975 shows the tree-ssa aliasing pass can be slow.
You can get the testcase from:
<http://www.math.purdue.edu/~lucier/GNATS/GNATS-12/_num.i.gz>
Here is the results from my build (yes with --disable-checking):
 tree PTA              :  35.74 (38%) usr   0.12 ( 1%) sys  37.63 (33%) wall


Note this testcase has a large number of computed gotos.

-- 
           Summary: [4.0 Regression] tree-ssa aliasing slow
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Keywords: compile-time-hog
          Severity: normal
          Priority: P2
         Component: tree-optimization
        AssignedTo: dnovillo at gcc dot gnu dot org
        ReportedBy: pinskia at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18419

[Bug tree-optimization/18419] [4.0 Regression] tree-ssa aliasing slow

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18419

[Bug tree-optimization/18419] [4.0 Regression] tree-ssa aliasing slow

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-11-10 15:36 -------
Most of the time is spent in bitmap_bit_p which is called from collect_points_to_info_for, maybe we 
should be using sbitmap instead.

Note collect_points_to_info is called from merge_pointed_to_info

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18419

[Bug tree-optimization/18419] [4.0 Regression] tree-ssa aliasing slow

[lucier at math dot purdue dot edu]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |lucier at math dot purdue
                   |                            |dot edu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18419

[Bug tree-optimization/18419] [4.0 Regression] tree-ssa aliasing slow

[steven at gcc dot gnu dot org]

------- Additional Comments From steven at gcc dot gnu dot org  2004-11-11 00:05 -------
Note that the number of computed gotos shouldn't matter much, we 
factor computed gotos to have a single common computed jump. 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18419

[Bug tree-optimization/18419] [4.0 Regression] tree-ssa aliasing slow

[steven at gcc dot gnu dot org]

------- Additional Comments From steven at gcc dot gnu dot org  2004-11-11 00:29 -------
You could try an sbitmap for "ai->ssa_names_visited" in tree-ssa-alias, 
and a pointer_set for "visited" in walk_use_def_chains_1. 
 

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-11-11 00:29:26
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18419

[Bug tree-optimization/18419] [4.0 Regression] tree-ssa aliasing slow

[steven at gcc dot gnu dot org]

------- Additional Comments From steven at gcc dot gnu dot org  2004-11-12 00:12 -------
http://gcc.gnu.org/ml/gcc-patches/2004-11/msg00920.html 
 
There's still room for speedups but PTA is off the map now. 
 

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18419