Re: Experimenting with tree inliner parameter settings

[Gerald Pfeifer <pfeifer at dbai dot tuwien dot ac dot at>]

On Mon, 5 May 2003, Steven Bosscher wrote:
> That does not look good, all but one are much slower :-(
>
> Would you know what the numbers look like with 3.2.3?

Please find below new results, including GCC 3.2.2.  Again, running times
in seconds.

                     |   GCC 3.2.2   | |   original    |   --params    |
---------------------+---------------+ +---------------+---------------+
      STRATCOMP1-ALL |  22.72 (0.01) | |  20.07 (0.04) |  18.62 (0.04) |
   STRATCOMP-770.2-Q |   0.85 (0.03) | |   0.87 (0.00) |   0.90 (0.00) |
               2QBF1 |  20.60 (0.01) | |  21.49 (0.05) |  22.30 (0.03) |
          PRIMEIMPL2 |  11.78 (0.01) | |  10.45 (0.01) |  10.74 (0.03) |
            ANCESTOR |   9.20 (0.01) | |  10.56 (0.05) |  10.82 (0.02) |
       3COL-SIMPLEX1 |   7.28 (0.02) | |   7.74 (0.02) |   8.00 (0.03) |
        3COL-LADDER1 |  41.31 (0.13) | |  49.72 (0.10) |  52.39 (0.36) |
      3COL-N-LADDER1 |  21.20 (0.00) | |  25.84 (0.11) |  26.12 (0.04) |
        3COL-RANDOM1 |  13.22 (0.01) | |  13.70 (0.03) |  13.94 (0.04) |
          HP-RANDOM1 |  15.31 (0.03) | |  17.99 (0.07) |  18.40 (0.06) |
       HAMCYCLE-FREE |   1.68 (0.00) | |   1.80 (0.00) |   1.88 (0.00) |
             DECOMP2 |  19.84 (0.04) | |  21.02 (0.06) |  22.49 (0.09) |
        BW-P4-Esra-a |  92.28 (0.04) | |  77.36 (0.09) |  78.56 (0.20) |
        BW-P5-nopush |   8.08 (0.01) | |   6.55 (0.02) |   6.64 (0.03) |
       BW-P5-pushbin |   6.91 (0.00) | |   6.24 (0.04) |   6.31 (0.01) |
     BW-P5-nopushbin |   2.16 (0.00) | |   2.03 (0.00) |   2.10 (0.02) |
              3SAT-1 |  36.17 (0.04) | |  33.38 (0.21) |  34.25 (0.06) |
   3SAT-1-CONSTRAINT |  20.20 (0.01) | |  19.94 (0.09) |  20.50 (0.05) |
        HANOI-Towers |   4.91 (0.01) | |   4.61 (0.02) |   5.03 (0.02) |
              RAMSEY |   8.89 (0.01) | |   9.28 (0.00) |   9.78 (0.03) |
             CRISTAL |  11.94 (0.02) | |  13.34 (0.03) |  13.70 (0.08) |
             HANOI-K |  38.91 (0.03) | |  40.25 (0.21) |  41.83 (0.11) |
           21-QUEENS |  10.37 (0.01) | |   9.34 (0.06) |   9.44 (0.03) |
   MSTDir[V=13,A=40] |  18.47 (0.00) | |  19.74 (0.00) |  20.22 (0.01) |
   MSTDir[V=15,A=40] |  18.58 (0.01) | |  19.97 (0.05) |  20.49 (0.04) |
 MSTUndir[V=13,A=40] |  10.23 (0.01) | |  10.81 (0.01) |  11.11 (0.01) |
         TIMETABLING |  11.26 (0.01) | |  11.12 (0.05) |  12.00 (0.02) |
---------------------+---------------+ +---------------+---------------+

> We're papering over troublesome algorithms in the back end, obviously
> the correct fix would be to get rid of those, but who's working on
> that??

CodeSourcery in the not too far future (as far as I understood) and also
the tree-ssa folks?

Gerald
-- 
Gerald "Jerry"   pfeifer@dbai.tuwien.ac.at   http://www.pfeifer.com/gerald/