This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Big differences on SpecFP results for gcc and icc
- From: Sebastian Pop <sebastian dot pop at cri dot ensmp dot fr>
- To: Scott Robert Ladd <scott dot ladd at coyotegulch dot com>
- Cc: "Menezes, Evandro" <evandro dot menezes at amd dot com>, Robert Dewar <dewar at adacore dot com>, Uros Bizjak <uros at kss-loka dot si>, gcc at gcc dot gnu dot org
- Date: Mon, 13 Jun 2005 20:33:03 +0200
- Subject: Re: Big differences on SpecFP results for gcc and icc
- References: <84EA05E2CA77634C82730353CBE3A84302AFBD80@SAUSEXMB1.amd.com> <42ADB691.5080703@coyotegulch.com>
Scott Robert Ladd wrote:
>
> My conclusion is the composite switches like -O2 are good only for
> general-purpose code. Anyone explicitly interested in squeezing out the
> most performance needs to do analysis and use application-specific switches.
>
Probably this situation is created by the fact that code
transformations are applied based on heuristic functions that were
written, well, in a very artisanal way.
http://www-osl.cs.umass.edu/~cavazos/thesis.ps describes how to
automatize the tuning of heuristic functions based on dynamic feedback
and pattern matching. The only difficulty remaining to the charge of
the compiler constructor is to effectively characterize the
fingerprints of programs with respect to a given transformation. Lots
of engineering, and cool speedups...
Sebastian