Based on your description, I can't think of a better (more "21st century")
way to do this.  A less intensive locking strategy (one lock for several
updates) would probably yield more impedes.  Are you doing entry-level
locking?  How is data locality -- would some detail repacks help?  What kind
of wait states are you seeing for this process?  Is it waiting for disc I/O?
Impedes?

-----Original Message-----
From: Cecile Chi [mailto:[log in to unmask]]
Sent: Monday, November 26, 2001 7:59 AM
To: [log in to unmask]
Subject: Re: 7.0 Performance hit?


Some additional analysis shows that our long-running MRP job is spending
most of its time in a subroutine where it does a lot of FLOCK - FREADDIR -
FWRITEDIR - FUNLOCK processing.  The long run time might be caused
by the impact of the "large files" code added in 6.5 and 7.0.
Perhaps a design done in the 70's for a Series 33 isn't the best design for
an application running on an N-class machine under 7.0.

Cecile Chi

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

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