Re: Why C is much slower than Fortran

Erik Corry <erik@arbat.com> wrote:
}> The only true solution for the alias problem is going to come from hardware:
}I saw a rather nice solution on Usenet a few years ago.
}All functions where optimisations are impossible due to C's liberal
}aliasing rules are automatically compiled in two versions, one where
}Fortran aliasing is assumed and one where it isn't. When you call the
}function the compiler calls the right one according to whether it can
}prove that arguments are unaliased. ...


}Has this been tried and failed?


      I wonder would it be sophisticated enough to tell the
      difference between something like this:


                void func(int * a, int * b, int n) {
                      int i;
                      for (i = 0; i < n; ++i) a[i] += b[i];
                }


      called with:


                a = malloc(100 * sizeof(int));
                b = malloc(100 * sizeof(int));
                      ...
                func(a, b, 100);


      and with:


                b = malloc(150 * sizeof(int));
                a = b + 50;
                func(a, b, 100);


John
--
John Hascall, Software Engr, Acropolis Project Manager, ISU Computation Center
http://www.cc.iastate.edu/staff/systems/john/index.html <=- the usual crud

Return to the comp.compilers page.
Search the comp.compilers archives again.