This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug tree-optimization/21883] [4.1 Regression] jump threading causing excessive code duplication



------- Comment #4 from steven at gcc dot gnu dot org  2005-10-07 21:32 -------
The patch has almost no effect except for -Os.  For SPEC binaries the 
effect of the patch is not exactly shocking on AMD64 at least: No effect
at all on compile time, no effect on performance, and almost no effect
on code size either:

param = value of --param max-jump-thread-duplication-insns=...
INT = total size of SPECint binaries at -O2 with this param
FP = total size of SPECfp binaries
value   INT     FP
1000    4353510 3349018
500     4353510 3349018
200     4353510 3349018
100     4352614 3348570
75      4352198 3347322
50      4351302 3346874
40      4351046 3346746
30      4350374 3346394
20      4350310 3346170
10      4350150 3341434
5       4346342 3340826
0       4345126 3341034

I don't have time to update the patch for recent changes and do any
further testing (e.g. CSiBE) to get this accepted, so unassigning.


-- 

steven at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|steven at gcc dot gnu dot   |unassigned at gcc dot gnu
                   |org                         |dot org
             Status|ASSIGNED                    |NEW


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21883


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]