HP3000-L Archives

February 2006, Week 4

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"Dave Powell, MMfab" <[log in to unmask]>
Reply To:
Dave Powell, MMfab
Date:
Tue, 28 Feb 2006 11:50:55 -0800
Content-Type:
text/plain
Parts/Attachments:
text/plain (31 lines)
Thanks to all who replied before.  I am up & running using Editor to extract a
file that then gets compiled.  Works like a charm and I plan to have a new
version of my HP2RTF command file available as soon as I pretty it up a bit.

The performance difference is.... well, for a dinky little 79-line input file,
the new version is about 3 times faster even 1st time, when it compiles part
of itself, and about 10 times faster thereafter.  Yes, compiled code is sooooo
much faster than interpreted that we can run Editor, the Cobol compiler and
the linker, and still read and write 79 records in a third the total time,
with lots of initialization stuff still interpreted.

And that result was stacking the deck in favor of the old version.  Stacking
the deck in favor of the new version it comes in a bit over 320 times faster,
on a 6000 line report that needs time-consuming conversion of codes for
italics, boldface, etc.  If I had the patience to test the old version on
bigger reports I could probably get slightly better numbers to report.

Extracting & compiling part of a command-file turns out almost trivially easy.
Seems like it might be a good trick for anyone else who has a really slow
command-file they want to distribute in one piece.


Kinda makes me feel like an idiot for spending so much time back in 2004
performance-tweaking the old all-MPE version.  Although I got the orig version
up & running pretty quickly, I spent FAR longer later trying to speed up the
detail processing in MPE than it took me this time to replace that part with
Cobol.

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

ATOM RSS1 RSS2