Jan Hubicka - Re: Some GCC 4.1 benchmarks (Re: Thoughts on LLVM and LTO) (original) (raw)

This is the mail archive of the gcc@gcc.gnu.orgmailing list for the GCC project.

• From: Jan Hubicka
• To: Scott Robert Ladd
• Cc: Jan Hubicka , Benjamin Kosnik , Daniel Berlin , rth at redhat dot com, gdr at integrable-solutions dot net, gcc at gcc dot gnu dot org
• Date: Tue, 22 Nov 2005 22:05:01 +0100
• Subject: Re: Some GCC 4.1 benchmarks (Re: Thoughts on LLVM and LTO)
• References: <200511221817.jAMIH2Co014676@porkchop.devel.redhat.com> <m3psos4k19.fsf@uniton.integrable-solutions.net> <1132685232.3076.132.camel@linux.site> <20051122184957.GA18372@redhat.com> <1132685583.3076.135.camel@linux.site> <20051122130700.090c4564.bkoz@redhat.com> <20051122200539.GB4648@atrey.karlin.mff.cuni.cz> <43837D1B.6040400@coyotegulch.com>

Jan Hubicka wrote:

I should note that comparison to ICC is not quite fair since it lacks Opteron tunning...

I think you may be comparing oranges to tangerines -- not as bad as apples and oranges, but still potentially an invalid comparison.

In my experience the extra registers of the Opteron provide a significant benefit; GCC has an unfair advantage if ICC only generated code for the small set of x86 registers.

Forgot to mention, all the tests was x86-64, so ICC used extra registers too. Also to clarify, the C++ benchmarks was done with -O3 only.

Honza

..Scott

• References:
  • Re: Thoughts on LLVM and LTO
    * From: Benjamin Kosnik
  • Re: Thoughts on LLVM and LTO
    * From: Gabriel Dos Reis
  • Re: Thoughts on LLVM and LTO
    * From: Daniel Berlin
  • Re: Thoughts on LLVM and LTO
    * From: Richard Henderson
  • Re: Thoughts on LLVM and LTO
    * From: Daniel Berlin
  • Re: Thoughts on LLVM and LTO
    * From: Benjamin Kosnik
  • Some GCC 4.1 benchmarks (Re: Thoughts on LLVM and LTO)
    * From: Jan Hubicka
  • Re: Some GCC 4.1 benchmarks (Re: Thoughts on LLVM and LTO)
    * From: Scott Robert Ladd