SPECFP95 comparison: 2.95.3 vs 3.0-20010525 on SPARC

[Dan Nicolaescu]

dewar@gnat.com writes:

  > <<Eh.  True, but one could also argue it wasn't being "published"
  > either.  This list is more of a forum for developers than a
  > vehicle for marketriods.
  > >>
  > 
  > I really think you should check the license agreement here ...
  > I don't see any exception in it about "developers" vs "marketroids"
  > 
  > besides which, as a developer, I would like to see the report!

We had a campus wide power failure here yesterday, and the machine
where I run this benchmark is not up yet (it's not maintained by our
support staff, but by myself and a colleague, and we haven't gotten
around turning on the machine). I'll post the complete report when we
do that. Sorry for the delay. 

(Meanwhile, the only thing that has any significance that is missing from what
I posted are 2 lines in each report that are the SPECfp scores for the
base and peak configurations. The scores are computed as geometric
means of the individual scores ie the numbers in the 4th column, after
the ===== line)