Re: gfortran 4.3.2 about 3x slower than g77 for scientific application

[Steve Kargl <sgk at troutmask dot apl dot washington dot edu>]

On Wed, May 13, 2009 at 10:07:20PM +0200, Julian King wrote:
> -march=native makes negligible difference to the running time.
> 
> I agree with you about the profiling being weird in terms of the
> number of calls. This is to the same data though, and the outputs of
> the program are the same. I just wonder if gprof is not catching all
> the calls somehow?
> 
> -Wall produces a large number of warnings, but all of them relate to
> 1) nonconforming tab characteres
> 2) unused dummy variables
> 3) unused variables.
> 

I was hoping that -Wall would find issues involving mixed
mode math, which can have an obvious influence on floating
point code.  Things like

   double precision x
   integer i
   x = 0.d0
   do i = 1, 100
      x = x + 0.1  ! 0.1 isn't exactly representable.
   end do

BTW,    dexpf is not a glibc function afaict.
BTW**2, the code at the nabble.com url is insufficient to 
        help you with your problem.

-- 
Steve