minnie.tuhs.org/unixen.tar/unixen/PDP-11/Distributions/other/Venix/RX50.notes/text0005.txt

Text File | 1996-02-17 | 3.6 KB | 78 lines

In article <[e-mail]> [e-mail] (Andy Hakim) writes:
>[e-mail] (Charles Lasner) writes:
>:
>: Where can I obtain Teledisk?  Does it require any low-level preformatting
>: of the media on the target?
>:
>I checked the ftp site "wuarchive.wustle.edu", it is in directory
>"mirrors2/msdos/dskutl" as file "teled212.zip".
>
>Nope, this one does not require any low level preformatting.  It's possible
>to stick in a brand new ds/dd and let it go.  Incidently, teledisk is
>made by the same people, Sydex, who make 22disk.
>
>-andy

Sydex also makes RAINDOS.

I suspect that teledisk will only make sector-compatible descendents though,
so if I have a specially layed-out version of a diskette (such as 2:1
interleave or staggered, etc.) the descendent will lose that aspect of
optimization, and will instead become "vanilla" RX50 format in the
case of RX50 diskette.

The point is that certain software, especially for DECmates not specifically
geared to CP/M-80, and *any* bootable DECmate diskette (including CP/M-80) the
format used in stock RX50 layout is non-optimal.  There are different
requirements for different specific applications, but just as on PC's, the
use of non-interleaved non-staggered disks can be demonstrated to be
inferior to a variant in terms of sector ordering at the low format level.

Rainbow MS-DOS disks have an implied software interleave of 2:1 for the
FAT area, and 1:1 in the rest; this is in software, so the standard disk
layout should be maintained, except that the *stagger* is not taken into
account.  Thus, like a PC, Rainbow MS-DOS disks should be formatted with a
stagger of 2 per track.  Thus track 1 is layed out 1,2,3,4,5,6,7,8,9,10 and
track 2 is layed out 9,10,1,2,3,4,5,6,7,8.  When the disk seeks from track 1
to track 2, it will thus miss 9 and 10, but immediately find 1.  Were the
stagger not there, it would miss 1 and 2, and reject 3,4,5,6,7,8,9 while
waiting for 1 to come around again.  Thus staggering relieves rotational
latency.

For the DECmate, there are two additional problems:

All bootable diskettes require the logically sequential reading of tracks
78, 79 in the order 1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10.  But the RX
interface of the DECmate can't perform 1:1 interleave ever, so this is
anti-optimal not only in stagger but more importantly in interleave.  Thus,
this area of the disk should be formatted with an interleave of 2:1 as well
as a stagger of 2.  Thus the disk is layed out:

track 78: 1,6,2,7,3,8,4,9,5,10
track 79: 5,10,1,6,2,7,3,8,4,9

This restriction is based on the ROM routines that read in this area in linear
order.  This is mostly why all DECmates take so long to boot up!  Changing to
a better sector order will chop seconds out of the boot time.

Further, all systems other than CP/M-80 require some form of help, mostly
applying the stagger that helps the Rainbow as well (again other than CP/M).
For DECmate MS-DOS, tracks 0-3 should be in 1-1 interleave because the software
already maps the disk in 2:1 usage.  tracks 4-79 should be formatted 2:1
interleave to help out the RX interface when the Rainbow-optimal ordering
is invoked (similar to the DECmate ROM access, and just as inefficient on
a DECmate).

OS/278 does a software 2:1 interleave, so the only help needed is a
disk-wide stagger factor of 2.

Note that RT-11 and all other -11-oriented disks should use stock format only
because this superior software maps all sectors to include both the 2-1
interleave and stagger of 2 already.

So, if Teledisk is a *really* good utility, it won't disturb the format's
stagger and interleave as it copies the disks!

cjl