Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!olivea!samsung!zaphod.mps.ohio-state.edu!ncar!gatech!purdue!ccncsu!longs.LANCE.ColoState.EDU!kk881595
From: kk881595@longs.LANCE.ColoState.EDU (kevin knappmiller)
Newsgroups: comp.sys.dec
Subject: Summary of f77 on DEC 5000
Keywords: FORTRAN DEC5000 MIPS
Message-ID: <11315@ccncsu.ColoState.EDU>
Date: 20 Nov 90 19:39:51 GMT
Sender: news@ccncsu.ColoState.EDU
Reply-To: kk881595@longs.LANCE.ColoState.Edu
Distribution: na
Organization: Engineering College, Colorado State University
Lines: 22

Thanks for all the recommendations.

It seems that the version 2.1 compiler from MIPS, also sold as the
version 2 compiler from DEC, is the best there is right now. 
Unfortunately it is what we have and I am still not happy with it. 
There was a comment that there is a hyper optimizer being used at Rice. 
If anyone has any specifics on this I would appreciate comments.

The only explanation for the SGI running faster was that memory caching
was implemented in a manner which the compiler was better able to take
advantage of.
At least this was the only explanation postulated which I was unable to
rule out.

The real result though, is that if one wants speedy code on the DEC
5000, one should program in C, not FORTRAN.  I will concetrate on
running my long FORTRAN codes on a CRAY YMP.

Kevin Knappmiller
Department of Mechanical Engineering
Colorado State University
Fort Collins, CO  80523