This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Fix PR 30735 and 31090: Change static memory partition heuristic
H. J. Lu wrote on 04/17/07 17:27:
> I got SPEC CPU 2006 numbers on x86-64. I tried
>
> 1. Revision 123724
> 2. Revision 123718 +
>
> http://gcc.gnu.org/ml/gcc-patches/2007-01/msg02335.html
What are you comparing here? Why are you not comparing revision X vs
revision X + patch?
>
> I found
>
> #2 vs. #1
> 410.bwaves 2.67857%
> 416.gamess -1.14286%
> 433.milc 0%
> 434.zeusmp -0.653595%
> 435.gromacs 0.107527%
> 436.cactusADM 0.943396%
> 437.leslie3d 15.8009%
> 444.namd -0.671141%
> 447.dealII -3.89105%
> 450.soplex 0%
> 453.povray 0%
> 454.calculix 6.57216%
> 459.GemsFDTD -6%
> 465.tonto 0%
> 470.lbm -0.699301%
> 481.wrf 1.92308%
> 482.sphinx3 0%
> SPECfp_base2006 0.75188%
What do these numbers mean? More context please. It seems to me that
the percentage is SPEC score delta from #1 to #2? Or are these SPEC
runtimes? Or SPEC compile times?
> Richard's simple patch no longer applies now after Diego's change is
> checked in. Richard, do you have a newer patch I can try?
Richard's patch is not a good idea wrt compile times. It essentially
reverts the partitioner behaviour to not partition most variables, which
is the same as trying a very high --param max-aliased-vops parameter.