home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
CP/M
/
CPM_CDROM.iso
/
mbug
/
mbug078.arc
/
TURBODOS.WRN
< prev
next >
Wrap
Text File
|
1979-12-31
|
5KB
|
90 lines
-----------------------------------------------------------------
===> If you run Turbo-DOS or Compupro Concurrent operating sys
tems, read the following message.
-----------------------------------------------------------------
CRUNCH and UNCRunch versions 2.x are one of a handful of util-
ities which support multi-sector I/O. "Multi-sector I/O" is a
feature of CPM+ (3.0) & MP/M which can provide significantly
improved speed performance when performing disk I/O. The CRUNCH
programs automatically decide whether to use multi-sector or
conventional I/O each time they run. This is accomplished by
doing a "Get System Version" call. If "2.2" is returned, meaning
the program is being run in a CP/M 2.2 type environment, conven-
tional I/O is automatically used. Similarly, if "3.0" is re-
turned, multi-sector I/O is used. So what's the problem?
Certain "compatible" operating systems, which return "3.0" as
their System Version, do not in fact support multi-sector I/O, at
least not correctly. SLR Systems are the people put out the VERY
high speed Z80ASM assembler. It, too, uses multi-sector I/O
where applicable. The following are excerpted from the Z80ASM+
Release 1.00 Notes:
-----------------------------------------------------------------
TurboDOS
"Since Multi-sector I/O does not handle EOF conditions on reads,
Multi-sector I/O sequential must be disabled under TurboDOS.
Software 2000 says nobody uses Multi-sector I/O so they may or
may not ever fix it. You are the one that sent them money, maybe
they'll listen to you!"
Compupro Concurrent
"This is another system that as of this writing does not support
Multi-sector even though it claims MP/M compatibility. Oddly
enough, if your source files are in Ram-disk, Multi-sector I/O
works, but on hard disk, forget it. This is another 'nobody'
uses it case. I wonder if nobody uses these performance-improv-
ing features because on most commercial systems they don't work?
Again, maybe they will listen to those who sent them money..."
-----------------------------------------------------------------
Like the Z80ASM, CRUNCH v2.1 provides a means by which multi-
sector I/O can be inhibited, regardless of the result returned by
the Get System Version call. Although this feature was also in
v2.0, it did require patching a single byte in each program. The
information was contained in the v2.0 documentation (though the
operating systems in question were not singled out by name). The
majority of the [sometimes vocal] minority of TurboDOS users were
apparently not aware of this deficiency. All reported TurboDOS
problems did disappear when these patches were made, however.
CRUNCH v2.1 requires no knowledge by the user of how to patch a
byte. It comes with a simple installation program, CRINSTAL,
which will ask you four simple questions; one of those questions
is whether or not you are running one of these systems. Run
CRINSTAL as described in CRINSTAL.DOC, answer the questions, and
there should be no problem.
-----------------------------------------------------------------
The Silver Lining: An update.
Maybe enough people did call. Software 2000 has announced that
they have in fact fixed the multi-sector I/O problems (as well as
other problems) as of the most recent release of TurboDOS, namely
v1.43. TurboDOS expert Jonathan Lodden confirms that CRUNCH v2.0
runs fine on v1.43, without any override patches. So.. IF you
HAVE TurboDOS 1.43, you may answer "NO" to the TurboDOS question.
Then you will receive the full speed benefits of multi-sector I/O
(which John calls "significant" under TurboDOS).
-----------------------------------------------------------------
NOTE: You can tell if CRUNCH and UNCRunch are using multi-sector
by observing the arrow between the filenames. If you see a
" ===> " (hi-power arrow), multi-sector I/O is being used.
Otherwise you will see the more standard " ---> " arrow.
If anyone has any updated information on Compupro Concurrent, or
other information on this subject, I would be pleased to hear
from you so that the information could be included in a future
release of this document.
- Steven Greenberg
201-670-8724 (voice) eves,wknds
■("¼φ*¿φ"ñφ├Q▀: