Re: ACCPTHSIZ perfomance - *MAX1TB vs *MAX4GB

Dave,

Thanks for response.  Just to be clear, I don't plan to do new development
defining some files at *MAX1TB and others at *MAX4GB.  All new files we create
use the default *MAX1TB.  I was more concerned about the situation where I have
a PF with 27 existing LFs all with *MAX4GB and I create a new LF with *MAX1TB.
The manual indicates that this will cause performance problems.  If this is
true, I want to change all my LF's to *MAX1TB immediately.  If not, it becomes a
lower priority project.

Phil

   Date: Thu, 03 Jan 2002 18:58:49 -0500
   From: David & Eileen Keck <bstars@optonline.net>
   Subject: Re: ACCPTHSIZ perfomance - *MAX1TB vs *MAX4GB
   To: midrange-l@midrange.com
   Reply-To: midrange-l@midrange.com

   A couple years back the firm I was with was testing JDE OneWorld, which was
   ODBC based.  Due to multi-minute response times, possible performance
   adjustments were examined ... *MAX1TB was one of them.  After changing the
   files related to the test (always all LF for a given PF) performance
   improved dramatically, though response time never became acceptable.  After
   reading up on ACCPTHSIZ I changed all files in the system on a Sunday using
   a PDM user option (thousands of files, some over 100,000,000 records, about
   70 Gig overall).  OS/400 uses a different algorithm to access records when
   using MAX1TB vs *Max4GB.  That algorithm is usually but not always faster.
   You asked about mixing *MAXxxx methods for LF's relating a common PF ... I
   can't see any reason to do that.
   -Dave K.