Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!spool.mu.edu!munnari.oz.au!metro!seagoon.newcastle.edu.au!cc.newcastle.edu.au!cegrw From: cegrw@cc.newcastle.edu.au Newsgroups: comp.sys.mac.programmer Subject: FORTRAN compiler summary Message-ID: <1991Mar8.145226.9698@cc.newcastle.edu.au> Date: 8 Mar 91 04:52:26 GMT Organization: University of Newcastle, AUSTRALIA Lines: 36 The general consensus of FORTRAN compilers is that 1. Absoft's MacFORTRAN II is about 20-30% faster on real codes (as opposed to benchmarks) than V2.1 of Language Systems FORTRAN. Some small tight loops may optimise better up to 50%. However, anocdotal evidence says that compile times are significantly longer for MACFORTRAN II (up to 10 times slower). With nooptimisation selected MACFORTRAN II was still faster but the speed of compilation differential was smaller. Language Systems themselves admitted to this differential but said that the next version will incorporate more optimisation. No comments on the size of executable code. 2. No significant comments came in comparing the TTY emulating windows that both provide but some suggested that Language Systems FORTRAN was easier to use and extend. 3. The VAX compatability of Language Systems FORTRAN was seen to be a superior point. Both compilers were seen to be bug free (unlike ABSoft's stand alone MACFortran/020 V2.4 which one reply described as a bug-ridden-port-from-hell which from my own experience is ungenerous but fairly true). 4. MacFORTRAN II is considerably more expensive than Language Systems and can not be bought without MPW. 5. No comments were made on SADE compatability of either compiler. 6. In Australia at least, technical support for both compilers is unsatisfactory. The general concensus was that Language Systems was the more useful compiler and better value though with the advent MacFORTRAN II the situation is no longer as clear cut as before. Dr Garry Willgoose Dept of Civil Engineering and Surveying University of Newcastle, NSW, 2308 AUSTRALIA.