Re: crash in 1.0.3a fortran compilation

[Jeffrey A Law <law at cygnus dot com>]

  In message <rzqk953lamo.fsf@djlvig.dl.ac.uk>you write:
  > A little caution: I've never seen SPEC but comments someone sent me
  > after reporting a problem with it suggest that at least some spec92
  > Fortran is quite broken (not unexpected...).  The bit in question
  > seemed to have been extensively fixed in spec95, but...  You can
  > probably make it work as intended with appropriate g77 options, but
  > you'll take a performance hit which may be misleading.
First I've heard of it.  But then again spec92 is quite old.

I'd love to switch to spec95, but there's licensing issues that would
have to be worked out.


  > There are some shorter free Fortran benchmarks we could deal with (as
  > well as the longer Livermore one Tim Prince as talked about) if
  > there's any interest.  Some of them also have C counterparts, which
  > might be relevant.
Yup.  Also note that much of the code which appears in spec is also
available from other sources for free.  The trick is you have to find
input codes and build stuff to verify that you're getting the correct
results :-)

jeff