HP3000-L Archives

August 1996, Week 3

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:
Goetz Neumann <[log in to unmask]>
Reply To:
Date:
Tue, 20 Aug 1996 19:31:11 +0200
Content-Type:
text/plain
Parts/Attachments:
text/plain (27 lines)
Mike Liening wrote:
>
> Hi all,
> We've recently migrated from our old series 70 to a 937 and I have a
> question regarding compatibility mode programs.  Our main application is
> written in compiled HP/Basic and uses *a lot* of SPL calls.  Looking at
> Glance, obviously we have a high percentage of Compatibility Mode
> applications running.  When I OCTCOMP the programs, the CM drops some, but
> then CM/NM Switches go way up (because of the SPL calls).  We don't
> currently have the Business/Basic compiler, nor do we have SPLash
> (although we are looking into both).  Question is... Is it best to leave
> the programs in CM to cut down on the switches, or should I OCTCOMP them
> to cut down on the CM?  Which (CM or Switches) is more demanding of MPE?
 
That is a difficult question (I am sure Stan might have input as well),
but generally the savings of OCTCOMPed CM programs (or user SLs) should
be much higher than the effort for additional switching.
OTOH, I think your performance tool will simply report more switches
per second, because the CM part of your code is executed faster, when
OCTed.
 
From 6 years of support experience I remember only one example where
OCTing a CM program made things worse; that example was the CM PASCAL
compiler.
 
Goetz.

ATOM RSS1 RSS2