From owner-freebsd-hackers Thu Jun 22 22:33:42 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id WAA22794 for hackers-outgoing; Thu, 22 Jun 1995 22:33:42 -0700 Received: from crh.cl.msu.edu (crh.cl.msu.edu [35.8.1.24]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id WAA22788 for ; Thu, 22 Jun 1995 22:33:41 -0700 Received: (from henrich@localhost) by crh.cl.msu.edu (8.6.11/8.6.9) id BAA09033 for freebsd-hackers@freebsd.org; Fri, 23 Jun 1995 01:34:31 -0400 From: Charles Henrich Message-Id: <199506230534.BAA09033@crh.cl.msu.edu> Subject: Poor program load time To: freebsd-hackers@freebsd.org Date: Fri, 23 Jun 1995 01:34:31 -0400 (EDT) X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 1223 Sender: hackers-owner@freebsd.org Precedence: bulk I have very often exprienced cases where I will attempt to run a program (be it ls or xterm) where for no apparent reason it takes seconds to load and execute. This is on a P100 w/ 32mb ram and 1gb SCSI disk (but It occurs on all 2.0.5-R systems). My guess is the problem is with the merged vm cache code. As my cache is always full of file data, when I attempt to execute a program there is some sort of contention while the system determines just what the hell it should throw out to get me up and running. AIX has a similar sort of merged cache but program execution is virtually instantaneous. One thing I've noticed with AIX is that when a program is first run it is started with a "higher" priority and as time goes on the priority is lowered. Now, could FreeBSD do the same thing to give "new" processes a kick in the butt so to the end user response time improves? Now all the above is based on supposition on a topic I know virtually nothing about, so If I sound like a rambling idiot please correct me, but explain why the load times are so damn long some times please! -Crh Charles Henrich Michigan State University henrich@crh.cl.msu.edu http://rs560.msu.edu/~henrich/