develooper Front page | perl.perl5.porters | Postings from April 2001

Re: [ID 20010425.012] VMS Perl pipes 2x slower on Perl 5.6.1 compared to Perl 5.5.3

From:
lane
Date:
April 26, 2001 05:39
Subject:
Re: [ID 20010425.012] VMS Perl pipes 2x slower on Perl 5.6.1 compared to Perl 5.5.3
Message ID:
010426084008.8ff15@DUPHY4.Physics.Drexel.Edu
Richard Stooke <Richard.Stooke@megasys.com> writes:

> My timing tests of Perl 5.6.1 show it be twice as slow in piping
> IO from a subprocess as Perl 5.5.3.

I'm not suprised at all; there's a lot more processing of piped data
in the recent Perls.  As Craig Berry notes, you'll also get a big
performance hit from not having a findable VMSPIPE.COM file, but
unless your config is messed up I doubt if lacking VMSPIPE is the
problem.

From looking at your script, the message size isn't large so I doubt
that mailbox size is a big issue.

As for what we get for performance hit, the answer is functionality.
Before the piping changes, we often had the test suite "hang" in
multiple places because of piping deadlocks and timing races.

Here's a message with a summary of some of the things the piping changes
were meant to address

  http://www.xray.mpe.mpg.de/mailing-lists/vmsperl/2000-03/msg00114.html

For completeness, the surrounding discussions in
    http://www.xray.mpe.mpg.de/mailing-lists/vmsperl/2000-03/threads.html
    http://www.xray.mpe.mpg.de/mailing-lists/vmsperl/2000-04/threads.html
are also informative.

And near the bottom of the web page:
    http://www.crinoid.com/perl560.htmlx
you'll see a section: VMS Perl "Pipe Torture Tests"

Give 'em a try ;)
--
 Drexel University       \V                    --Chuck Lane
======]---------->--------*------------<-------[===========
     (215) 895-1545     _/ \  Particle Physics
FAX: (215) 895-5934     /\ /~~~~~~~~~~~        lane@duphy4.physics.drexel.edu



nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About