|Date: ||Mon, 17 Jan 2005 16:33:03 -0500|
|Reply-To: ||"Alexander J. Shackman" <shackman@PSYPHW.PSYCH.WISC.EDU>|
|Sender: ||"SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>|
|From: ||"Alexander J. Shackman" <shackman@PSYPHW.PSYCH.WISC.EDU>|
|Subject: ||Re: spss performance issues (workspace, cache, execute)|
thanks for the additional advice raynald, joram & jon.
i tried raynald's more compact code and it works well -- although it adds
~20 seconds to the run time.
For the record, I've appended Jon's general advice (emailed directly to me)
for running lengthy syntax. Because I'm using Raynald's very compact code,
little is gained by doing as he suggests, but I imagine it would help a
good deal with longer code.
From: Peck, Jon <firstname.lastname@example.org>
Subject: RE: [SPSSX-L] spss performance issues (workspace, cache, execute)
My preliminary test shows that DO REPEAT is probably the fastest approach
to this problem, but there are a few other generally useful things to do
with a very long syntax file.
- turn off journaling (Edit/Options/General). Because the journal file is
volatile and tends to provoke virus checkers a lot, it can have huge
overhead. (We reworked this mechanism in SPSS 13 to mostly solve this
- do not have "echo commands to the log" turned on (Edit/Options/Viewer)
- run the file using INCLUDE (or INSERT in SPSS 13) rather than opening it
in the syntax window and running it. We improved performance in this area
in SPSS 13.0.1, but even with that, there is a fair amount of messaging
between the syntax window and the SPSS Processor that using INCLUDE avoids.