Compiling with and without -march=native: benchmarks

[Corrado]

Dear Octave friends,

here is the benchmark comparison between Octave 322 on Kubuntu 904 64 bit 
compiled from source with and without the --march=native flag set.

It seem that with the new gcc the flag -march-native yields higher 
performances. 

What do you think?

PS: I believe recompiling lapack and blas with -march=native may lead to even 
higher performances, but I have found lapack / blas recompilation rather 
troublesome.

Regards 

COMPILED WITH -O2 -MARCH=NATIVE
octave:1> benchmark
Octave benchmark version bm 2.3
Speed of octave 3.2.2 on x86_64-unknown-linux-gnu
  Pentium II 350 MHz has score 1.0, higher numbers are better
Matrix inversion (LAPACK)        14.98 +/- 0.1% (252 runs)
Schur decomposition (LAPACK)     54.52 +/- 0.4% (322 runs)
Differential equation (LSODE)    40.81 +/- 0.2% (980 runs)
Fourier transforms (FFTPACK)     23.09 +/- 0.4% (609 runs)
for loop                         75.20 +/- 3.6% (1365 runs)
-- Performance index (bm 2.3): 36

COMPILED WITH -O2
octave:4> benchmark
Octave benchmark version bm 2.3
Speed of octave 3.2.2 on x86_64-unknown-linux-gnu
  Pentium II 350 MHz has score 1.0, higher numbers are better
Matrix inversion (LAPACK)        15.11 +/- 0.3% (259 runs)
Schur decomposition (LAPACK)     54.36 +/- 0.2% (322 runs)
Differential equation (LSODE)    38.54 +/- 0.2% (910 runs)
Fourier transforms (FFTPACK)     21.55 +/- 2.8% (567 runs)
for loop                         63.70 +/- 1.0% (1183 runs)
-- Performance index (bm 2.3): 34




-- 
Corrado Topi

Global Climate Change & Biodiversity Indicators
Area 18,Department of Biology
University of York, York, YO10 5YW, UK
Phone: + 44 (0) 1904 328645, E-mail: address@hidden