Path: utzoo!attcan!uunet!husc6!mailrus!ames!killer!mic!d25001
From: d25001@mic.UUCP (Carrington Dixon)
Newsgroups: comp.lang.fortran
Subject: PL/I performance
Summary: Not always worse than others
Message-ID: <163@mic.UUCP>
Date: 17 Sep 88 17:00:30 GMT
References: <13587@mimsy.UUCP> <3698@lanl.gov> <423@quintus.UUCP>
Reply-To: d25001@mic.UUCP (Carrington Dixon)
Organization: Micro Net
Lines: 19

In article <423@quintus.UUCP> ok@quintus.UUCP (Richard A. O'Keefe) writes:
>
>PL/I is not a language that springs to mind when one thinks "high
>performance".  Was it that people didn't try hard enough, or are there
>fast *full* PL/I systems, or is Fortran 8X doomed by the resemblance?

    Several months ago I did some benchmarking of compilers on an IBM 3090
system.  In terms of computation speed (scalar only, of course) IBM's PL/I
compiled code was _very_ close to that of VS FORTRAN.  The SAS/C compiler was
worse by a factor of two and the C compiler that IBM was then selling --
they've just announced a newer (and better?) one -- was down by something
like a factor of four.
    I/O performance was again pretty close for PL/I and FORTRAN with PL/I
in the lead.  SAS/C trailed again by a factor of two and the IBM C by
an abysmal factor of ten or so.


Carrington Dixon
UUCP: { convex, killer }!mic!d25001