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


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

RE: i386 backend


> > it can be nice to turn it into kind of benchmark and send
> > me. The I can
> > investigate whats going wrong.

Hi all

I have found similar behaviour with gcc - it normally generates slower and
bigger executables that VC++ or Borland C++ (and even LCC) on Windows and
Dynix native C compiler on Dynix/ptx i386. However, as others have pointed
out, vague complaints do not serve any purpose.

What I would like to know is:

a) Whether there are any benchmarking suites available which one can run on
different compilers and compare the outputs.
b) Whether there are code snippets available which test different
optimizations.

If someone can point out these, I'll be more than happy to run them and send
assembly outputs and benchmark results to the GCC team for further analysis.

Biswa.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]