Re: Les options, ça change tout :

Personne n'a essayé l'option -mfpmath=sse de gcc?
la différence sur mon pentium 4m est flagrante, voila les résultats :
cc -O3 -march="pentium4" -pipe -funroll-all-loops -fomit-frame-pointer -ffast-math :
** **
** SciMark2 Numeric Benchmark, see http://math.nist.gov/scimark(...) **
** for details. (Results can be submitted to pozo@nist.gov) **
** **
Using 2.00 seconds min time per kenel.
Composite Score: 345.87
FFT Mflops: 200.95 (N=1024)
SOR Mflops: 243.92 (100 x 100)
MonteCarlo: Mflops: 80.61
Sparse matmult Mflops: 516.03 (N=1000, nz=5000)
LU Mflops: 687.83 (M=100, N=100)
et avec
cc -O3 -march="pentium4" -funroll-all-loops -fomit-frame-pointer -ffast-math -mfpmath=sse :
** **
** SciMark2 Numeric Benchmark, see http://math.nist.gov/scimark(...) **
** for details. (Results can be submitted to pozo@nist.gov) **
** **
Using 2.00 seconds min time per kenel.
Composite Score: 366.41
FFT Mflops: 219.13 (N=1024)
SOR Mflops: 268.38 (100 x 100)
MonteCarlo: Mflops: 70.09
Sparse matmult Mflops: 522.20 (N=1000, nz=5000)
LU Mflops: 752.25 (M=100, N=100)

cat /proc/cpuinfo :
processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 Mobile CPU 1.80GHz
stepping : 4
cpu MHz : 1798.831
cache size : 512 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm
bogomips : 3555.32