gfortran speed benchmarks

[Jerry DeLisle]

Giovanni Bajo wrote:
> Jerry DeLisle <jvdelisle@verizon.net> wrote:
> 
> 
>>What do you all think?  With Pauls patches.
>>
>>$ gfc -O3 -static -march=pentium4 -fmax-stack-var-size=1000000
>>-funroll-loops TEST_FPU.f90
>>                              total = 24.8 sec
>>
>>$ ifort -O3 TEST_FPU.f90
>>                              total = 24.3 sec
> 
> 
> 
> I think that the command line stinks. What's the results just with gfc -O3?
> I don't think it's fair if we get to the performance with some
> secret --super-performance. For what I know, your secret --super-performance
> may work well for this testcase and bad for others.

No secrets here.

You missed the point entirely.  These tests show that gfortran performance is in 
a reasonably good range and that it is approximate to a good commercial 
compiler.  The compiler flags are not relevant.  Users can tweak flags all they 
want.  There is nothing "standard" or magic about -O3.  It just means the 
compiler developers selected some set of optimizations based on some judgement 
and called it -O3.  Intel -O3 is not the same as gcc -O3.  These are two very 
different products.

The information I presented is intended for developers who know what the flags 
mean and don't read into it any more than is intended.  It is simply a measure 
of relative progress from before the patch and after the patch.

-- Jerry