Info-IBMPC@WSMR-SIMTEL20.ARMY.MIL (04/20/89)
Info-IBMPC Digest Wed, 19 Apr 89 Volume 89 : Issue 41
Today's Editor:
Gregory Hicks - Chinhae Korea <COMFLEACT@Taegu-EMH1.army.mil>
Today's Topics:
ARCMaster version 4.23 uploaded to Simtel20
Borland and the 8087
CADKEY Libraries
DRIVPARM problem
Windows Printer Driver for Okidata Microline 84
Borland TDebugger and '386 Problems
Borland Floating Point Emulation
386 max question
AT Questions
PCjr Hard Disk -- Does anyone know how?
FTP Access to WSMR-Simtel20.army.mil (3 msgs)
Disk Drive Information
Information Regarding Frolic Software
Multitask/user environment for MSDOS available from Simtel20
Possible to write to a "Write Protected" Disk
Turbo problems
Simtel20 Address, Benchmark and 360K/1.2M Problems
----------------------------------------------------------------------
Date: Wed, 12 Apr 1989 13:38 MDT
From: Keith Petersen <w8sdz@WSMR-SIMTEL20.ARMY.MIL>
Subject: ARCMaster version 4.23 uploaded to Simtel20
I have uploaded ARCMaster 4.23 to Simtel20:
Filename Type Bytes CRC
Directory PD1:<MSDOS.ARC-LBR>
AM423.ARC BINARY 152552 D220H
This is a menu-driven shell for use with ARC/PAK/PKPAK/ZIP, etc.
Be sure to specify a non-used directory as its work directory. I hear it
does DEL *.* in the work directory when it finishes. No problem if
installed correctly. Read the manual.
--Keith Petersen
Maintainer of Simtel20's CP/M, MSDOS, and MISC archives
Internet: w8sdz@wsmr-simtel20.army.mil [26.2.0.74]
Uucp: {ames,decwrl,harvard,rutgers,ucbvax,uunet}!wsmr-simtel20.army.mil!w8sdz
------------------------------
Date: Wed, 12 Apr 89 23:15 EDT
From: <SEMICON%WATSCI%watmta.UWaterloo.CA@CORNELLC.cit.cornell.edu>
Subject: Borland and the 8087
>I wonder if anyone else has noticed this ? I was testig relative
>performance with and without an 8087 using Borland Turbo C v1.5 and
>ProFortran. The test I used was to model a system based on the old
...
>The Turbo C code was ca. 10K smaller when compiled for the 8087 (due to
>the lack of emulator code) than the emulated executable. The ProFortran
>execuatble was 5K smaller when using 8087 than when emulating the
>floating point. However, when run, there was no difference in execution
>time with Turbo C between emulated floating point and 8087 code. With
>ProFortran there was a tenfold difference (the programs were slightly
>different between the 2 languages so they cannot be compared directly,
>although TurboC was faster than ProFortran without 8087 code but
>ProFortran was faster with it).
>Can anybody explain this? Does Borland have incredibly fast 8087
>emulation or lousy 8087 support or a mixture of the 2 ?
Lets deal with this one question at a time. First the emulation. The
Borland compiler replaces the first part of floating point instructions
with an interrupt. At startup this interrupt is initialized to point to
the emulator. Now the fun part 8{). When the emulation routines are
initialized (at startup) the code checks for the presence of an 80x87 and
if it is there it sets a flag. Now when the interupt is called and that
flag is set the emulator replaces the instruction with the 'real' floating
point code. At this point the resulting code is identical (most of the
time) to that produced for the non emulation version. The upshot of all
this is that unless you only use an instruction once or twice you will not
note a speed difference between having or not having an emulator.
Now the speed of the 8087 code. I don't know much about the code
Prospero generates but I suspect it uses 64 bits at most for the floating
point (probably only 32 bits if you are using REAL*4's). Borland, on the
other hand uses 80 bits for all its internal calculations. I supect the
speed difference you are seeing is due to the difference in precision
between the two compilers.
Robert Adsett <SEMICON@WATSCI.BITNET>
<SEMICON@WATSCI.UWaterloo.ca>
Dept. of Phys.
Univ. of Waterloo
Waterloo Ont. Canada
------------------------------
Date: 31 Mar 1989 09:51-CST
From: "410 BMW/SCX--KI Sawyer AFB MI" <SAC.2001CS-XP@E.ISI.EDU>
Subject: CADKEY Libraries
We are looking for any libraries of CADKEY patterns which may be available
for downloading. We are specifically interested in parts relating to
communications, but will take most anything we can.
We have been told that some AUTO-CAD libraries can also be used by CADKEY,
but haven't confirmed that yet.
Any assistance would be greatly appreciated.
Thanks,
TSgt Michael Barnes
------------------------------
Date: Mon, 10 Apr 89 12:43:46 BST
From: Bob Eager <rde%ukc.ac.uk@NSS.Cs.Ucl.AC.UK>
Subject: DRIVPARM problem
I solved the DRIVPARM problem in PC-DOS 3.3 in a different way that
requires NO patches. It's weird, but it seems to work!
After the DRIVPARM=, but before the parameters, add THREE Control-A
characters (yes, ASCII code 1). I have no idea why this works...
------------------------------
Date: Mon, 06 Mar 89 16:05:50 EST
From: SSROB%ECUVM1.bitnet@cunyvm.cuny.edu
Subject: Windows Printer Driver for Okidata Microline 84 step 2
Does anyone know of a printer driver for Microsoft Windows that works with
an Okidata Microline 84 Step 2 printer. The printer is not IBM Graphics
Compatibl e. Better yet does anyone know the format of the Windows
Printer Driver files or how to create one. Thanks in advance.
Rob L. Hudson ___ ___ SSROB@ECUVM1.BITNET
Systems Programmer |__ | | | ( 919 ) 757 - 6401
East Carolina University |___ |___ |__| Greenville, NC 27858
------------------------------
Date: Fri, 7 Apr 89 16:53:57 N
From: molliet@elde.epfl.ch
Subject: Borland TDebugger and '386 problems
I recently purchased a copy of Turbo Debugger for my PC. I tried to run
this software using the 8086-virtual mode offered by the 386 processor but
the system is always trashed by the 386-specific version (TD386.EXE) of
Turbo Debugger.
In my config.sys file, I put only the following commands:
FILES=20
BUFFERS=20
LASTDRIVE=D
DEVICE=ANSI.SYS
DEVICE=TDH386.SYS (Borland driver for Turbo Debugger in virtual mode)
As soon as TD386 is trying to load the Turbo Debugger program (TD.EXE), I
get the following message on the screen (and the system hangs completely):
Unexpected interrupt 06 at 9F00:000001FE (flag=000330012) error code = NONE
EAX=00160400 EBX=000A0004 ECX=0000F6F9 EDX=000101F7
ESI=00000049 EDI=00000000 EBP=000000AE stk=1062:0000049E stack:0000 0000 0000
0000 0000 0F09 C000 0000 0000 0000 0000 0000 0000
My machine is a Hewlett-Packard RS/16 which is a 386 machine. It is
equipped with 2Meg of RAM, a 387 and an EGA Display.
I am running under DOS 3.20 with
ROM BIOS C.02.01 RS/16
RAM BIOS B.01.03 (which are Phoenix and HP BIOS)
The Norton SysInfo reports the followind data:
Computer Name: IBM AT
Operating System: DOS 3.20
Built-in BIOS dated: Monday, February 1, 1988
Main Processor: Intel 80386 Serial Ports: 1
Co-Processor: Intel 80387 Parallel Ports: 1
Video Display Adapter: Enhanced Graphics (EGA), 256 K-bytes
Current Video Mode: Text, 80 x 25 Color
Available Disk Drives: 4, A: - D:
DOS reports 636 K-bytes of memory:
143 K-bytes used by DOS and resident programs
493 K-bytes available for application programs
A search for active memory finds:
640 K-bytes main memory (at hex 0000-A000)
32 K-bytes display memory (at hex B800-C000)
64 K-bytes extra memory (at hex E000-F000)
ROM-BIOS Extensions are found at hex paragraphs: C000
The same program runs perfectly on a Compaq Deskpro 386 which is a rather
similar machine.
Has anyone experienced the same problem using the Turbo Debugger program
in 8086-virtual mode. Any answer will be welcommed.
Thank you in advance.
Philippe Molliet
Electronics Labs
Swiss Federal Institute of Technology at Lausanne
Switzerland
E-MAIL: molliet@elde.EPFL.CH
------------------------------
Date: Mon, 10 Apr 89 10:57:00 PDT
From: <madler%Hamlet.Bitnet@cunyvm.cuny.edu> (Mark Adler)
Subject: Borland Floating Point Emulation
The Borland floating point emulation code will use an 80x87 if one is
there. In this case, both the emulation and 80x87 versions of the code
will run at almost the same speed (almost, because the emulation version
replaces floating point calls with instructions as they occur).
For testing purposes, you can force the emulation version to actually use
emulation by setting the environment variable "87" to "N". For example:
C>set 87=N
before executing the program, will keep the emulation code from looking
for an 80x87. You can also "set 87=Y" to force the code to think there is
an 80x87 there, even if there isn't one! The program will crash, of
course, if there isn't an 80x87 in this case. I have no idea why they
provide this option. You can "set 87=" to return things to normal. The
87 environment variable is checked in Turbo C 1.5 and later.
Mark Adler
bitnet: madler@hamlet
arpa: madler@hamlet.caltech.edu
------------------------------
Date: Wed, 12 Apr 89 10:14 N
From: <MMKOISTINEN%FINKUO.bitnet@cunyvm.cuny.edu>
Subject: 386 max question
Could someone tell me what is difference between 386MAX and 386MAX
Professional. I only know that 386MAX Professional is twice as expensive
as 386MAX.
thanx,
Mika Koistinen * Kraftwerk : I program my home computer
Lataajanpolku 1A6 * Dire Straits: That ain't working thats the way you do it
70460 KUOPIO * decnet: kylk::mmkoistinen BITNET: MMKOISTI@FINKUO
FINLAND * internet: mmkoistinen%kylk@opmvax.kpo.FI
------------------------------
Date: Wed, 12 Apr 89 11:19 EST
From: JOHN MARKETON <MARKETON%SUNRISE.bitnet@cunyvm.cuny.edu>
Subject: AT Questions
I am working on a program that will be making use of the protected mode
on an AT type computer. I have several questions about the protected
mode, and the books I have don't answer those questions. If you have any
info, please email it to me ASAP. Thanks in advance.
1 - When in protected mode, where happens to the video memory? I need to
know what happens to HGC, MDA, CGA, EGA, and VGA.
2 - There appears to be a conflict between BIOS INT 10 (the video
interrupt) and the NPX's INT 10 (NPX fault). Can I still use the BIOS
video services provided through INT 10? If both the video services and
the NPX share INT 10, how can I service NPX faults, but pass video
services to the BIOS?
3 - Is it possible to copy the BIOS routines into memory (regular or
extended) to speed up BIOS operation? This assumes that shadow RAM isn't
available, and that the copy is a movestring operation.
4 - How are the DMA chips used (both for memory-memory, memory-disk and
disk-memory transfers)?
5 - What are the differences between the 8250 and the 16450 UARTS?
Thanks,
John Marketon
MARKETON @ SUNRISE (Bitnet)
SUNRISE.ACS.SYR.EDU (Internet)
128.230.1.1 ( " )
------------------------------
Date: Wed, 12 Apr 89 21:47 EST
From: Soh Chai <YCHEAH%hampvms.bitnet@cunyvm.cuny.edu>
Subject: PCjr Hard Disk -- Does anyone know how?
Help! I have a IBM PCjr with 640K RAM (upgraded from 128K IBM SideCar by
PC Enterprise), IBM RGB Monitor, IBM Parellel SideCar, Enchaced keyboard
(not Chicklets), new Fujitsu 360K floppy disk drive (2 made for IBM Qume
drives died).
With my current hardware, and running DOS 3.10, I manage to run
WordPerfect 4.2, Lotus 123 2.01, dBase III+, Turbo Pascal 3.0 etc with
minor fuss. However, I want to stick a hard disk on this system so that I
can totally eliminate all the fuss of disk swapping.
Does anybody out there knows how to, or better yet, have added a hard disk
on to a PCjr. I would very much prefer to do it myself since most
companies charge >UDSS $700 for a 20 mb 65 ns Seagate with controller.
Thanks.
------------------------------
Date: Tue, 11 Apr 89 04:26:46 MDT
From: Gregory Hicks <GHICKS@WSMR-SIMTEL20.ARMY.MIL>
Subject: FTP Access to WSMR-Simtel20.army.mil
Have you tried the following commands to your USER FTP process?
Pick one of the following:
TYPE LOCAL (lets your ftp proc and the server work out the protocol)
TYPE LOCAL 8 ( -do- )
TYPE TENEX (Tells the server you want 36 bit words sent in 1K blocks without
transformation)
That may help... If you can capture the output of your user FTP process
to a file, I'd like to look at your 'help' file from the FTP process.
Perhaps I can suggest something. I don't have a VAX, or I'd be able to
give more help.
Best,
Gregory Hicks
------------------------------
Date: Mon, 10 Apr 89 21:10:31 MDT
From: Gregory Hicks <GHICKS@WSMR-SIMTEL20.ARMY.MIL>
Subject: FTP Access to WSMR-Simtel20.army.mil
You need to set the FTP server to the proper mode. In most cases, the
server @WSMR defaults to TEXT. Most of the files are binary.
Try a "TYPE LOCAL" or "TYPE LOCAL 8" or "TYPE TENEX" and see what
happens... Give the command to your user FTP process...
Best,
Gregory Hicks
------------------------------
Date: 22 March 1989 18:32:01 CST
From: "Michael J. Steiner" <U23405%UICVM.bitnet@cunyvm.cuny.edu>
Subject: A problem with FTP access to SIMTEL20...
For the past few days, whenever I have tried to FTP to SIMTEL20 (from
UICVM, which is at Chicago, Illinois), I receive a "foreign host rejected
the open attempt" message. Does anyone know if something is wrong with
SIMTEL20 that causes this? It is important to me to be able to get files
from SIMTEL20 via FTP because LISTSERV at RPICICGE is too restricted (a
maximum of three files per day or 100K, whichever is smaller).
Any help, hints would be appreciated.
Michael Steiner
Email: U23405@UICVM.BITNET
[Part of the answer to this problem may be that the WSMR-Simtel20.army.mil
network address changed from [26.0.0.74] to [26.2.0.74] ...]
------------------------------
Date: 12 Apr 89 12:22:59 PDT (Wednesday)
From: "George_C._Burkitt.ElSegundo"@Xerox.COM
Subject: Disk Drive Information
>Can the z-157 controller (which is part of the parent board) handle this
many floppys?
Can't help with the Zenith; I'd guess you can handle two drives with no
problem. If the extra drives don't work on the 'daisy - chained cable,
you can get a floppy bus extender card to allow hookup of the others.
Try it first, though, with just the cables.
>Will the fact that I don't have a physical B drive....
The jumpers apply only to floppy disk drives. The hard disk drives are
controlled by a separate controller.
>Just exactly what is "daisy chaining"
The floppy controller cable (usually a 34 conductor ribbon cable) runs
from the controller (a header type of connector) to each drive in one long
run, an edge - card connector fastened onto the ribbon cable for each
drive; spaced far enough apart to allow assembly into the box(es). Remove
the resistor pack (usually socketed) on all drives except the one at the
far end of the cable from the controller. In 5 1/4" drives the resistor
pack is usually a 14 or 16 pin DIP (I forget which); in the 3 1/2" drives
it may be a SIP. Set the Drive Select (DS) switch or jumper on each drive
to different DS numbers...DS1 on the A drive, DS2 on the next, etc.
(Sometimes the lowest DS number is 0, not 1!!). Daisy - chaining means
that all signals to and from all floppy drives are sent on the one cable;
the controller can 'select' the desired drive electrically. At the drive
end this selection is mirrored by means of the DS switches / jumpers.
There are signals in the ribbon cable to select the DS lines separately,
so that only one drive will be selected at a time. You have to insure
that only one drive uses each DS assignment. The DS control is either a
DIP Switch or jumper block, and is usually fairly near the connector and
resistor pack.
------------------------------
Date: Sun, 9 Apr 89 14:23:38 PDT
From: Mark_Wong@mtsg.ubc.ca
Subject: Information Regarding Frolic Software
Does anyone out there have info or a reference source for a software
product called Frolic - written by Datafile Rightline. The software is
supposed to help manage a filing system.
Mark E. Wong
Access Data Consultants Limited
7845 Allman Street
BURNABY, BC
CANADA V5E 2A9
USERLW47@MTSG.UBC.CA
USERLW47@UBCMTSG (BITNET)
------------------------------
Date: Sun, 9 Apr 89 15:16:54 EDT
From: "Leslie C. Brown" <lbrown@TBD.BRL.MIL>
Subject: Multitask/user environment for MSDOS available from Simtel20
I've uploaded the following file to Simtel20:
pd1:<msdos.sysutl>VMIX151.ARC Multitask/multiuser environment for MSDOS
VMiX is a multitasking and multiuser environment for IBM PC's and (PC, XT,
AT, 386AT) compatibles. VMiX is not for the IBM PS/2's.
Version 1.51 upgrades 'vw.exe', a pull-down menu utility that adds ease of
use to the VMiX shell.
The environment has a slight UNIX flavor in the implementation of the user
control shell and multitasking. The VMiX command shell offers built-in
help, or at the option of the user, the use of pull-down menus.
VMiX multitasks at the PC console (up to 4 windows) and/or serially
connected modem, PC or dumb terminal(s). VMiX can task switch
applications from the low 640K of conventional memory to extended
80286/386 high memory.
Applications that do not write directly to the video, like DBASE, TURBO
PASCAL, PROCOMM, WSTAR, WORDPERFECT, OPUS BBS, or BASIC will display
correctly at a terminal.
VMiX co-resides with DOS 2.1 through 3.3 The VMiX 'exec' command will pass
any user desired command to the DOS control shell(s) 'command.com'. The
VMiX shell can also be configured to look exactly like a DOS shell, in
which case the 'exec' command becomes superfluous.
Les
------------------------------
Date: 10 Apr 89 17:44:00 CST
From: zielke@physics.rice.edu
Subject: Possible to write to a "Write Protected" Disk
In reference to the sure fire cure for viris problems using a bootable
disk in drive A which is "write protected". This write protection is
performed in software at some level. It is possible "At least on a Real
IBM-AT 6mhz, first rom revision" to write directly to the disk and bypass
the write protect mechanism. I do not know how it was done but I know
that it can be done, I ran across someone who had written this code so as
to be able to write on disks with no notch cut in them...
David M. Zielke
ARPA==> Zielke@Physics.Rice.Edu
Zielke@128.42.9.23
MaBell==> 713-527-8101 ext. 4018 work
713-666-2982 home
US Snail==> David M. Zielke
7490 Brompton #110
Houston, Tx 77025
------------------------------
Date: Tue, 11 Apr 89 12:42 EST
From: BCA%PSUARCH.BITNET@CUNYVM.CUNY.EDU
Subject: Turbo problems
I have been having a lot of hangup problems with my XT turbo system that I
put together mysef..
The problem is this : When I am at 4.77 Mhz everything works beautifully..
But when I go to Turbo (10Mhz), and leave it there for about 20 minutes,
the computer will lock up at any given time. Sometimes it will go for
hours before locking up... I especially have had lock-ups when in Procomm
Plus. I can tell when things are about to go bizerk, because when I am in
the dialing directory (ALT-D), some of the border-line characters are
missing.
One other thing- before I got my Hard Drive fixed and put it back in, I
was running on 2 360K'ers. Everything worked fine, including turbo at 10
Mhz...
Strange, Huh?!
Here's my set-up, if this could trigger a solution from anyone...
CGA graphics, 640K (10 rows of 150ns NEC 64k ram), Seagate ST-225 with
Western Digital controller, 2 floppies, 1 1.44 meg 3.5" drive, mouse
(which I leave un-driven), 150 W PS, and a printer.
I have noticed some wierd things that might be hints:
1. With a chkdsk run, I am told that I only have 639K. Other programs
tell me the same thing.
2. When I get the missing border characters in Procomm Plus, I can turn
off the turbo (back to 4.77), and everything works fine again...
My BIOS is a Phoenix 2.52 and my processor is an Intel 8088-1..
Anyone have any solutions or similar problems??
Thanks... Barry Allyn BCA@PSUARCH.BITNET
------------------------------
Date: Thu, 19 Apr 89 16:43:00 MEX
From: Mario Camou Riveroll <EM302723%VMTECMEX.BITNET@CUNYVM.CUNY.EDU>
Subject: Simtel20 Address, Benchmark and 360K/1.2M Problems
What is the complete address of the Simtel20 archives?
[The host address is WSMR-SIMTEL20.ARMY.MIL [26.0.0.74]. This address
will change about 20 March to [26.2.0.74].] gph]
Subject: IBM PC benchmarks source code
Where can I get the standard benchmark source code? (Dhrystone,
Whetstone, etc.)
[Whetstone sources are in PD1:<MSDOS.SYSUTL> as follows:
"PD1:","<MSDOS.SYSUTL>","WHETST.ANS",1,1405,7,851230
"PD1:","<MSDOS.SYSUTL>","WHETST.FOR",8,4012,7,840129
gph]
Subject: 360K vs 1.2Mb problems
I've been following the discussion on the incompatiblity problems between
360K floppies formatted in DSDD & 1.2M drives. Does anyone know whether
there are also problems with 720K & 1.44M drives?
Mario Camou
EM302723@VMTECMEX
#include <disclaimer.h>
------------------------------
End of Info-IBMPC Digest
************************
-------