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 optimization/3041] [3.3/3.4 regression] Stepanov Abstraction Penalty regression for powerpc-ibm-aix


------- Additional Comments From pinskia at gcc dot gnu dot org  2003-12-07 07:23 -------
On the mainline or 3.3.2(20030908), I get with -O2:

test      absolute   additions      ratio with
number    time       per second     test0

 0        0.00sec    3333333.33M         1.00
 1        0.00sec    3333333.33M         1.00
 2        0.00sec    3333333.33M         1.00
 3        0.00sec    3333333.33M         1.00
 4        0.00sec    3333333.33M         1.00
 5        0.00sec    3333333.33M         1.00
 6        0.00sec    3333333.33M         1.00
 7        0.00sec    3333333.33M         1.00
 8        0.00sec    3333333.33M         1.00
 9        0.00sec    3333333.33M         1.00
10        0.00sec    3333333.33M         1.00
11        0.00sec    3333333.33M         1.00
12        0.00sec    3333333.33M         1.00
mean:     0.00sec    3333333.33M         1.00

Total absolute time: 0.00 sec

Abstraction Penalty: 1.00
While on the 2.95.3, I get this with -O2:

test      absolute   additions      ratio with
number    time       per second     test0

 0        0.00sec    2000000.00M         1.00
 1        0.00sec    1923076.92M         1.04
 2        0.00sec    3333333.33M         0.60
 3        0.00sec    1666666.67M         1.20
 4        0.00sec    3333333.33M         0.60
 5        0.00sec    1666666.67M         1.20
 6        0.00sec    3333333.33M         0.60
 7        0.00sec    1666666.67M         1.20
 8        0.00sec    3333333.33M         0.60
 9        0.00sec    1612903.23M         1.24
10        0.00sec    3333333.33M         0.60
11        0.00sec    1612903.23M         1.24
12        0.00sec    3333333.33M         0.60
mean:     0.00sec    2341349.82M         0.85

Total absolute time: 0.00 sec

Abstraction Penalty: 0.85

So every thing looks fixed.
In fact this was the same for 3.3 so it was fixed for 3.3 (yes, -O1 is has a different result).

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|tree-ssa                    |---


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=3041


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