Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!evax!cs4344af
From: cs4344af@evax.arl.utexas.edu (Fuzzy Fox)
Newsgroups: comp.sys.cbm
Subject: Re: 1571 writes slowly on side 2
Keywords: bam 1571 disk side two write slow
Message-ID: <1991Mar29.195031.16595@evax.arl.utexas.edu>
Date: 29 Mar 91 19:50:31 GMT
References: <1991Mar28.024743.16062@gn.ecn.purdue.edu> <46392@ut-emx.uucp>
Organization: Computer Science Engineering Univ. of Texas at Arlington
Lines: 22

In article <46392@ut-emx.uucp> andy@ccwf.cc.utexas.edu (Andrew Hackard) writes:
>If I remember correctly, the problem is that many of the tracks on side 2 are
>allocated by the BAM on side 1; when the 1571 writes a sector on side 2, it has
>to go back to side 1 to mark the block as used.  I don't know why that would
>make things so much slower...

You've got it right.  The problem is that there is only room in the 1571
RAM to stores one of the two BAM blocks.  Also, some of the info is
stored in BAM block 1 (side 1), and some in block 2 (side 2).  The fact
that they are on the same track, but different sides, helps a bit, but
the main problem is that the head must seek to track 18 over and over,
which really slows things down.

The newer ROMs simply cache the BAM changes in RAM (as is normally done
when side 1 is being allocated), so the seek to track 18 is not needed
until the file is closed.

-- 
David DeSimone, aka "Fuzzy Fox" on some networks.          /!/!
INET:    an207@cleveland.freenet.edu                      /  ..
Q-Link:  Fuzzy Fox                                        /   --*
Quote:   "Foxes are people too!  And vice versa."         /  ---