This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gcc 3.3 got faster.
- From: Peter Schmid <schmid at snake dot iap dot physik dot tu-darmstadt dot de>
- To: <toon at moene dot indiv dot nluug dot nl>
- Cc: <gcc at gcc dot gnu dot org>
- Date: Sun, 29 Sep 2002 21:09:52 +0200 (CEST)
- Subject: Re: gcc 3.3 got faster.
I made a similar observation for g++ on the i686-pc-linux-gnu target
running the speed test suite included in the FTensor-1.1pre20
distribution. gcc version 3.3 20020927 is faster that gcc version 3.3
20020915 in all cases, except for Tensor 1, confer the appended
results.
The setting for the CXXFLAGS was: -ftemplate-depth-1000
-Drestrict=__restrict__ -O3 -finline-functions -finline-limit-1000
-funroll-loops.
Hope this helps,
Peter Schmid
20020915 20020927
fast 1
real 0m4.991s real 0m3.108s
user 0m4.840s user 0m3.100s
sys 0m0.000s sys 0m0.000s
Tensor 1
real 0m3.583s real 0m5.346s *
user 0m3.580s user 0m5.250s *
sys 0m0.000s sys 0m0.040s *
fast 2
real 0m19.036s real 0m18.710s
user 0m18.890s user 0m18.200s
sys 0m0.010s sys 0m0.000s
Tensor 2
real 0m40.675s real 0m28.678s
user 0m40.370s user 0m28.450s
sys 0m0.000s sys 0m0.000s
fast 3
real 0m5.489s real 0m4.004s
user 0m5.440s user 0m3.960s
sys 0m0.000s sys 0m0.000s
Tensor 3
real 0m7.754s real 0m4.313s
user 0m7.680s user 0m4.310s
sys 0m0.000s sys 0m0.000s
fast 4
real 0m9.386s real 0m7.701s
user 0m9.300s user 0m9.770s
sys 0m0.010s sys 0m0.000s
Tensor 4
real 0m12.254s real 0m9.842s
user 0m12.170s user 0m9.770s
sys 0m0.000s sys 0m0.000s
fast 5
real 0m12.927s real 0m9.834s
user 0m12.820s user 0m9.770s
sys 0m0.000s sys 0m0.000s
Tensor 5
real 0m22.971s real 0m18.410s
user 0m22.840s user 0m18.280s
sys 0m0.000s sys 0m0.000s
fast 6
real 0m5.402s real 0m5.100s
user 0m5.360s user 0m5.050s
sys 0m0.000s sys 0m0.000s
Tensor 6
real 0m6.278s real 0m5.410s
user 0m6.200s user 0m5.360s
sys 0m0.000s sys 0m0.000s
fast 7
real 0m11.339s real 0m10.875s
user 0m11.260s user 0m10.710s
sys 0m0.000s sys 0m0.000s
Tensor 7
real 0m11.300s real 0m10.376s
user 0m11.230s user 0m10.290s
sys 0m0.000s sys 0m0.010s
fast 8
real 0m17.367s real 0m17.678s
user 0m17.240s user 0m17.570s
sys 0m0.000s sys 0m0.000s
Tensor 8
real 0m17.951s real 0m16.567s
user 0m17.750s user 0m16.430s
sys 0m0.000s sys 0m0.000s
fast 9
real 0m25.016s real 0m25.375s
user 0m24.810s user 0m25.210s
sys 0m0.000s sys 0m0.000s
Tensor 9
real 0m25.330s real 0m23.457s
user 0m25.140s user 0m23.310s
sys 0m0.000s sys 0m0.000s