Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site umcp-cs.UUCP
Path: utzoo!linus!decvax!genrad!panda!talcott!harvard!seismo!umcp-cs!chris
From: chris@umcp-cs.UUCP (Chris Torek)
Newsgroups: net.lang.f77
Subject: Re: Fast, reliable FORTRAN compilers for UNIX
Message-ID: <1101@umcp-cs.UUCP>
Date: Mon, 5-Aug-85 00:20:18 EDT
Article-I.D.: umcp-cs.1101
Posted: Mon Aug  5 00:20:18 1985
Date-Received: Wed, 7-Aug-85 01:46:59 EDT
References: <174@mdivax1.UUCP>
Organization: U of Maryland, Computer Science Dept., College Park, MD
Lines: 16

>It is my understanding that one of the main problems contributing
>to poor execution times is the nature of the floating point library
>-- ie: all floating point is double precision.

I'm not sure about the "all floating point is double precision" part
(not having worked on any of that myself), but slow library routines
are indeed one of the reasons for slow execution of Fortran routines
on 4.2BSD Vaxen.  There is a "native" math library available (-lnm)
which you can use if you dare (it is at least somewhat more buggy
than the conventional Fortran math library, I believe).

4.3BSD has a revised native math library for Vaxen, I hear.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 4251)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris@umcp-cs		ARPA:	chris@maryland