"Experimental" features in releases

[Toon Moene]

Paolo Bonzini wrote:
> 
>> We use to call this "benchmarketing"
>>
>> I don't see why GNU would want to do that for anything.
> 
> 
> Because (that's unfortunate, agreed) GCC does need some marketing. 
> Unfortunately people compare GCC with icc (or other compilers) using 
> SPEC,

Hmmm, that's not my experience (note that SWIM is Fortran and is listed 
as a primordial Weather Forecasting Benchmark - I could explain why, but 
then I have to produce quite a bit of TeX ;-)

What people in my environment do is compare ifc <whatever options they 
got from their coworkers> against gfortran -O2.

This will never fly, of course.

The first thing I teach them is to add -ffast-math -funroll-loops.

Oh, and last week taught *me* to also suggest -msse2 -mfpmath=sse for 
those using Intel processors (of course, I use Debian AMD64 testing, so 
for me that's the default :-)

-- 
Toon Moene - e-mail: toon@moene.indiv.nluug.nl - phone: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
A maintainer of GNU Fortran 95: http://gcc.gnu.org/fortran/
My next laptop will be pedal-powered