This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PING][PATCH/RFT] Fix PR middle-end/PR28690, modify swap_commutative_operands_p
- From: "H. J. Lu" <hjl at lucon dot org>
- To: Peter Bergner <bergner at vnet dot ibm dot com>
- Cc: "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>, Paolo Bonzini <paolo dot bonzini at lu dot unisi dot ch>
- Date: Sun, 6 May 2007 13:21:16 -0700
- Subject: Re: [PING][PATCH/RFT] Fix PR middle-end/PR28690, modify swap_commutative_operands_p
- References: <1177614080.20709.34.camel@otta> <20070428153118.GA515@lucon.org> <1177795225.20709.42.camel@otta> <20070429235551.GA12511@lucon.org> <1177935400.20709.44.camel@otta>
On Mon, Apr 30, 2007 at 07:16:39AM -0500, Peter Bergner wrote:
> On Sun, 2007-04-29 at 16:55 -0700, H. J. Lu wrote:
> > That is very interesting. On Linux/i686, I found
> >
> > 1. 252.eon worked with gcc 4.3 revision 123850 at -O2.
> > 2. 252.eon failed with gcc 4.3 revision 123850 at -O2 -ffast-math.
> > 3. 252.eon failed with gcc 4.3 revision 123850 plus your patch at -O2.
> > 3. 252.eon worked with gcc 4.3 revision 123850 plus your patch at
> > -O2 -ffast-math.
> >
> > It is hard to measure SPEC CPU 2K impact of your change on Linux/i686.
>
> Well, other than eon which clearly seems to have issues with or
> without the patch, what did the other benchmarks show?
Here are what I got on Core 2 Duo with -O2 -ffast-math:
1. SPEC CPU 2006 cores
(new - old)/old
400.perlbench 0.995025%
401.bzip2 0.613497%
403.gcc 0%
429.mcf 0%
445.gobmk 1.16279%
456.hmmer -0.775194%
458.sjeng 0.574713%
462.libquantum 0%
464.h264ref 0.393701%
471.omnetpp 0.877193%
473.astar 0%
483.xalancbmk 0.634921%
Est. SPECint_base2006 0.666667%
410.bwaves 0%
416.gamess 1.15607%
433.milc 0%
434.zeusmp -0.653595%
435.gromacs 0.865801%
436.cactusADM -0.943396%
437.leslie3d -0.972973%
444.namd -1.34228%
447.dealII -4.2471%
450.soplex 0%
453.povray 0.492611%
454.calculix -6.18557%
459.GemsFDTD 0%
465.tonto 0%
470.lbm 0%
481.wrf 0.961538%
482.sphinx3 0%
Est. SPECfp_base2006 -0.75188%
2. SPEC CPU 2000 cores
164.gzip 0.454545%
175.vpr -0.0541126%
176.gcc 1.2218%
181.mcf -0.60208%
186.crafty 0.388007%
197.parser 0.0712758%
252.eon -0.746269%
253.perlbmk -0.851064%
254.gap -1.5748%
255.vortex -0.233191%
256.bzip2 0.091659%
300.twolf -0.890631%
Est. SPECint_base2000 -0.226757%
168.wupwise -6.56276%
171.swim 0.844595%
172.mgrid 0.453258%
173.applu 2.65291%
177.mesa 0.916086%
178.galgel -1.08573%
179.art -4.50307%
183.equake 0.688765%
187.facerec 0.440852%
188.ammp -2.92006%
189.lucas 1.68684%
191.fma3d -0.949868%
200.sixtrack 1.64114%
301.apsi 0.877646%
Est. SPECfp_base2000 -0.463939%
On Core Duo, I got
1. SPEC CPU 2006 cores
400.perlbench 0%
401.bzip2 1.04167%
403.gcc 0.943396%
429.mcf 0%
445.gobmk 0%
456.hmmer 0.164745%
458.sjeng 0%
462.libquantum 0%
464.h264ref 0%
471.omnetpp 0.263852%
473.astar 0%
483.xalancbmk 0.146199%
Est. SPECint_base2006 0.217155%
410.bwaves 1.59744%
416.gamess 0.238095%
433.milc 0.544218%
434.zeusmp 0.293686%
435.gromacs -0.9375%
436.cactusADM -3.32717%
437.leslie3d -0.589391%
444.namd 1.1509%
447.dealII -2.32558%
450.soplex 0.443951%
453.povray 0.648649%
454.calculix 0.684932%
459.GemsFDTD 0.158228%
465.tonto 0.157729%
470.lbm -8.15508%
481.wrf 0.4158%
482.sphinx3 0%
Est. SPECfp_base2006 -0.584795%
2. SPEC CPU 2000 cores
164.gzip 0.671892%
175.vpr -0.0913242%
176.gcc 0%
181.mcf -0.789266%
186.crafty -0.556881%
197.parser 0.208333%
252.eon X
253.perlbmk 0.280308%
254.gap 0.45283%
255.vortex 0.227445%
256.bzip2 1.13529%
300.twolf -2.60764%
Est. SPECint_base2000 0.16247%
168.wupwise 0.695652%
171.swim 1.59516%
172.mgrid 1.76991%
173.applu 1.12486%
177.mesa 0.825593%
178.galgel -2.08248%
179.art -3.31384%
183.equake 0.566974%
187.facerec -0.28222%
188.ammp -5.73566%
189.lucas -0.0845309%
191.fma3d -4.3722%
200.sixtrack 1.01215%
301.apsi 0.413223%
Est. SPECfp_base2000 -0.574163%
H.J.