Z-80 coprocessor cards
Greegor
for IBM XT 8 bit ISA slots.
Did anybody ever compare them to each other?
Did anybody ever make a PCI car version?
While software emulation now kind of obsoletes them,
wouldn't they be a good basis for hardware interfacing
projects?
For example, it looks like all current add in cards
that would supposedly upgrade an older PC
to USB 2 would require at a minimum PCI slots
2.0. The problem is most older PCs that
would need the USB upgrade don't have PCI
slots that new! (Quite a catch!)
Has anybody ever hooked up USB port chips
to Z80 processors?
Which of the newer faster incarnations of the
Z-80 would be the most true to the Z-80
instructions and function except for a higher
clock rate? (Wouldn't a 3 GHz Z-80 be hilarious?)
Peter Dassow
> Has anybody ever hooked up USB port chips
> to Z80 processors?
Yes. See http://www.ez80sbc.com/ for example. Because a normal Z80 is
too slow and has not enough memory address space for a complex USB
software stack, interfacing must be done with the help of intelligent
USB device controller.
> Which of the newer faster incarnations of the
> Z-80 would be the most true to the Z-80
> instructions and function except for a higher
> clock rate? (Wouldn't a 3 GHz Z-80 be hilarious?)
A 3 GHz Z80 will make no sense, because of the limited memory address
space and also the limited instruction set compared to a modern CPU.
Also, as already stated from yourself, a software emulation will be very
fast, fast enough for almost all purposes.
And, you have always also the possibility to 'hardware emulate' the Z80
by implementing the CPU into an ASIC / FPGA, using VHDL & VERILOG
development environments.
Regards
Peter
CBFalconer
>
... snip ...
>
> A 3 GHz Z80 will make no sense, because of the limited memory
> address space and also the limited instruction set compared to
> a modern CPU. Also, as already stated from yourself, a software
> emulation will be very fast, fast enough for almost all purposes.
> And, you have always also the possibility to 'hardware emulate'
> the Z80 by implementing the CPU into an ASIC / FPGA, using VHDL
> & VERILOG development environments.
Simply emulate a Hitachi 64180, which I believe is identical to a
Zilog Z180. It has a 1 megabyte address range, and is fully Z80
compatible.
--
[mail]: Chuck F (cbfalconer at maineline dot net)
[page]: <http://cbfalconer.home.att.net>
Try the download section.
no....@no.uce.bellatlantic.net
wrote:
>Peter Dassow wrote:
>>
>... snip ...
>>
>> A 3 GHz Z80 will make no sense, because of the limited memory
>> address space and also the limited instruction set compared to
>> a modern CPU. Also, as already stated from yourself, a software
>> emulation will be very fast, fast enough for almost all purposes.
>> And, you have always also the possibility to 'hardware emulate'
>> the Z80 by implementing the CPU into an ASIC / FPGA, using VHDL
>> & VERILOG development environments.
>
>Simply emulate a Hitachi 64180, which I believe is identical to a
>Zilog Z180. It has a 1 megabyte address range, and is fully Z80
>compatible.
Or emulate a Z280 as that had a scatter/gather MMU that could address
16Mb. If one had time and the latest FPGAs you can get it to the
40-50mhz region or better. For an 8bitter that screaming.
Then again you can buy an Ez80 wich can have ethernet and USB.
Why would 3ghz be a waste because of limited address? there are many
tasks that speed is more important than size.
Allison
fru...@pacbell.net
...
> Or emulate a Z280 as that had a scatter/gather MMU that could address
> 16Mb. If one had time and the latest FPGAs you can get it to the
> 40-50mhz region or better. For an 8bitter that screaming.
My four year old Athlon system running straight-forward C code
emulates a Z80 at 125 MHz. With a faster CPU and careful assembly
language coding, this could be at least doubled. If you really wanted
to scream, use some form of JIT transcoding to native x86 code. The
size of the win, or loss, depends on the application.
...
> Why would 3ghz be a waste because of limited address? there are many
> tasks that speed is more important than size.
Agreed. A 3 GHz Z80 could emulate a 30 MHz 6502. And a 30 MHz 6502
could emulate a 1 MHz Z80.
Glen Herrmannsfeldt
(snip)
> Or emulate a Z280 as that had a scatter/gather MMU that could address
> 16Mb. If one had time and the latest FPGAs you can get it to the
> 40-50mhz region or better. For an 8bitter that screaming.
I am getting interested in doing this. There is an FPGA board
with a large (rated at 1M gates) FPGA, VGA, serial, PS/2
keyboard/mouse port, Flash memory (mostly for configuration data),
ethernet, and either SRAM or DRAM for about $120.
That should be big enough for some pretty good sized whole systems.
I would think anything with a 6502, 68xx, 8080, or Z80 could easily
fit in there. I am wondering about 68020 (Sun3), SPARC (Sun4),
or some other larger systems. Maybe even VAX.
It should be big enough for many retrocomputing projects.
-- glen
Peter Dassow
> On Tue, 25 Nov 2008 22:26:35 -0500, CBFalconer <cbfal...@yahoo.com>
> wrote:
>
> Then again you can buy an Ez80 wich can have ethernet and USB.
> Why would 3ghz be a waste because of limited address? there are many
> tasks that speed is more important than size.
If speed is more important, why not using an already existing and cheap
fast CPU like an Intel Core2Duo or if you really like a current ARM chip
? Believe me, I like the Z80 design and the possibilities, but stack
oriented programming languages are doomed with a Z80 (yes, there is also
an Index register which will be used for similar purposes, but compare
it for example with a 6502, THAT would be a much more stack oriented
CPU). And modern CPUs are just more "compatible" with modern stack
oriented programming languages, so why cripple your possibilities down ?
So like "if you are in Rome, do as the romans do" I guess you live
today, so why not using todays CPUs for todays purposes ?
Regards
Peter
--
* http://www.z80.eu/blog - a vintage computer blog
dott.Piergiorgio
>> On Tue, 25 Nov 2008 22:26:35 -0500, CBFalconer <cbfal...@yahoo.com>
>> [...]
>> Then again you can buy an Ez80 wich can have ethernet and USB.
> That's why I mentioned the EZ80SBC ;-)
This (the eZ80) can be the best solution; IIRC a 40Mhz eZ80 are
comparable to an (hypothetical) 200Mhz Z80; of course, one must use care
with timed Z80 code because of the different execution time of
instructions (IIRC mainly LDIR & LDDR).
IMHO this also implies that a theoretical s-100 eZ80 CPU board has much
more problems & issues than benefits.
Hope to not have written stupidities....
Best regards from Italy,
Dott. Piergiorgio.
no....@no.uce.bellatlantic.net
wrote:
>no....@no.uce.bellatlantic.net wrote:
>> On Tue, 25 Nov 2008 22:26:35 -0500, CBFalconer <cbfal...@yahoo.com>
>> wrote:
>>
>> [...]
>> Then again you can buy an Ez80 wich can have ethernet and USB.
>That's why I mentioned the EZ80SBC ;-)
>
>> Why would 3ghz be a waste because of limited address? there are many
>> tasks that speed is more important than size.
>
>If speed is more important, why not using an already existing and cheap
>fast CPU like an Intel Core2Duo or if you really like a current ARM chip
Likely thats what I'd do. Heck for 100$ or so you can get a gumstick
ARM that is plenty fast with IO.
The arguement was a super fast is useless and it's not. There are a
slew of lesser machines running at amazing speeds as single chips
that have less rom/ram than z80. On the other hand interfacing any
chip that runs that fast and building a board around it is
non-trivial. With that few here would attempt building around BGAs
and 6-10 layer boards.
>? Believe me, I like the Z80 design and the possibilities, but stack
>oriented programming languages are doomed with a Z80 (yes, there is also
>an Index register which will be used for similar purposes, but compare
>it for example with a 6502, THAT would be a much more stack oriented
>CPU). And modern CPUs are just more "compatible" with modern stack
>oriented programming languages, so why cripple your possibilities down ?
Stack oriented languages do fine on z80. While the 6502 is a very
good chip the address return stack is pitiful compared to z80.
Allison
Howard Harte
> On Wed, 26 Nov 2008 08:16:08 +0100, Peter Dassow <z8...@arcor.de>
> wrote:
>
> >> On Tue, 25 Nov 2008 22:26:35 -0500, CBFalconer <cbfalco...@yahoo.com>
I have two z80 FPGA designs I've been working on. One is a PCI card
that has a z80 core from opencores.org, and it runs at 33MHz. It has
32K of SRAM implemented in the FPGA, and busmasters across PCI to get
another 32k of RAM from the PC. This allows easy loading of code
because the memory can be shared between the PC and Z80. It also has
a bunch of I/o mapped peripherals like UART.
The other design works on the Xilinx Spartan 3E Starter Kit, and has
the same z80 core, a FLASH interface, UART, VGA and PS/2 keyboard
interface. The VGA interface and keyboard emulates the Vector Graphic
FlashWriter2, including correct font ROMs. I run the Vector Monitor
on it and it works well. It also has a memory management unit and an
SDRAM interface, but I have problems with the SDRAM. Aside from that,
and lack of a disk controller, it is a complete Vector system in FPGA.
If people have interest in working on it, I can post the source code
to the Opencores.org website.
-Howard
Dennis Boone
> the same z80 core, a FLASH interface, UART, VGA and PS/2 keyboard
> interface. The VGA interface and keyboard emulates the Vector Graphic
> FlashWriter2, including correct font ROMs. I run the Vector Monitor
> on it and it works well. It also has a memory management unit and an
> SDRAM interface, but I have problems with the SDRAM. Aside from that,
> and lack of a disk controller, it is a complete Vector system in FPGA.
> If people have interest in working on it, I can post the source code
> to the Opencores.org website.
That'd be sweet.
De
Dennis Boone
> today, so why not using todays CPUs for todays purposes ?
I'm always amused to see people in comp.os.cpm questioning those who
would tinker with old technology. :)
Question: The FX2 USB chip contains an 851, iirc. So is an 8051 "todays
product", or is the FX2 "old technology"?
Lots of applications exist in which older designs are more economical.
De
Paul Potter
> too slow and has not enough memory address space for a complex USB
> software stack, interfacing must be done with the help of intelligent
> USB device controller.
That looks great.
Greegor
> [...] On the other hand interfacing any chip
> that runs that fast and building a board
> around it is non-trivial. With that few
> here would attempt building around
Thanks for pointing that out.
Big issue but would it be unusual to run
a prototype at a lower speed than production version, if any?
OK so maybe the idea of a 3 GHz Z-80
was a bit obtuse, but it reminds me of
what the late Grace Hopper (USN) predicted about
having a massive number of microprocessors
each handling one pixel on a video image.
Has the idea of massive numbers of cheap
microprocessors, each handling a tiny function died off?
Has Hopper's prediction of massively distributed
processing died off in favor of 4 centralized processors
functioning as one? Is there an I/O bottleneck issue
which causes a diminishing return?
When I Googled ISA USB board I got a shock.
I had no idea that somebody actually makes
a board to accept ISA/EISA (AT) slot cards
and connect them to a brand new computer
by way of USB 2.0!
I laughed and suddenly no idea of mine seems quite as bizarre.
Stranger things have happened!
Some Chinese company makes licensed reincarnations
of the old NES video game and another outfit once produced
Commodore 64 on a chip! They were sold as one of those
cheap ""retro"" video games built into a joystick to connect to a TV.
Could we have a super cheap modern incarnation
of a Z-80 system to run the open source CP/M?
Would there be any market, say as a laptop?
Even the PC world did a notebook PC for less than
$100 didn't it? But they're not marketed in the US right?
Tarkin
> no.s... (Allison) wrote:
> > [...] On the other hand interfacing any chip
> > that runs that fast and building a board
> > around it is non-trivial. With that few
> > here would attempt building around
> > BGAs and 6-10 layer boards. [...]
>
> Thanks for pointing that out.
> Big issue but would it be unusual to run
> a prototype at a lower speed than production version, if any?
>
> OK so maybe the idea of a 3 GHz Z-80
> was a bit obtuse, but it reminds me of
> what the late Grace Hopper (USN) predicted about
> having a massive number of microprocessors
> each handling one pixel on a video image.
>
> Has the idea of massive numbers of cheap
> microprocessors, each handling a tiny function died off?
>
I don't believe so. See comp.lang.forth and/or Intellasys....
> Has Hopper's prediction of massively distributed
> processing died off in favor of 4 centralized processors
> functioning as one? Is there an I/O bottleneck issue
> which causes a diminishing return?
>
I don't understand the Intellasys architecture well enough
to comment, but it's something like that, I think. As well
as the parallel-programming paradigm.
But then, consider modern graphics chipsets... a very
parallel architecture, dedicated to doing one specific
subsystem, and doing it well...
Here's a linky on Microwulf, a multiboard, network-based
cluster:
http://www.clustermonkey.net//content/view/211/1/
I was looking for link to an article about some researchers
having built a custom-pc that had between 4 and 8
graphics cards, and then just used the GPUs as
parallel computing units. I couln't find it, but recent
news shows big companies getting in this, with
GPU-based supercomputers.
And from what I understand of x86 PC, that's been the
bottleneck for some years now. I have computers that
~ 1GHz. On the ones that still have a 100 or 133 MHz bus,
the I/O bottleneck is obvious. Even my dual-core (32 bit, though),
DDR-memoried laptop is hindered by bus speed and harddrive
access times (running Vista and having a low-end drive doesn't help).
> When I Googled ISA USB board I got a shock.
> I had no idea that somebody actually makes
> a board to accept ISA/EISA (AT) slot cards
> and connect them to a brand new computer
> by way of USB 2.0!
>
Hey, there are still COBOL programmers...
Industrial computing and process control still use
a lot of very old designs...what was the app or
market for what you saw? I know some S100,
STD, and VME bus parts, cards, cages, etc
are still being sold. Ditto for 8088, 8086, and
80386.
> I laughed and suddenly no idea of mine seems quite as bizarre.
>
> Stranger things have happened!
>
> Some Chinese company makes licensed reincarnations
> of the old NES video game and another outfit once produced
> Commodore 64 on a chip! They were sold as one of those
> cheap ""retro"" video games built into a joystick to connect to a TV.
>
Ahhh, the DTV, was it?
> Could we have a super cheap modern incarnation
> of a Z-80 system to run the open source CP/M?
> Would there be any market, say as a laptop?
>
Probably not a large one. IMHO, the consumer is
simply too conditioned for visuallly-driven computing.
The cli is (too) obtuse for many, many people. What's even
stranger is that even close friends that I have, who
worship efficiency for efficiencies' sake, resist learning
even keyboard shortcuts. You can have their mouse when
you pry it from their...etc.
> Even the PC world did a notebook PC for less than
> $100 didn't it? But they're not marketed in the US right?
I believe you are referring to the One Laptop Per Child
initiative (OLPC), which predictably got mired in bureaucracy
and commitee design (IIRC and IMHO). The price went from
100$ to 125$ at one point, I and I have no idea what has
happened to the project since.
A 100$ CP/M Z?80-based laptop would have a certain
appeal. But you'd have to sell quite a few to break even.
The main costs would be design, screen, board fab,
and assembly, in roughly that order, I think.
And to get beyond the hacker market segment, you'd
have to find a way to get some graphics on there. Perhaps
some sort of Tektronix video mode, separate from the OS.
Perhaps even pipe it out separate from the cli main screen.
Then, you could do plotting, maybe CAD/CAM, fractals
and such. With even such modest graphical capabilities,
you might capture some research/educational/scientific
marketshare.
Here's a crazy idea for you:
Why not fab 4 Z?80 cores + mem for each +
I/O on a single chip? Task-switching would be
'real; task-switching, in that, instead instead of
saving a bunch state information, the core just
goes into standby, and the next core gets the
I/O? With either shared memory or traditional
storage, you could edit a source file in one
core, assemble it on another, link it on
the third, and execute it on the fourth.
I'm thinking the memory for each core might
be something akin to cache memory on x86.
Feasible, or pipe dream? Probably the wrong
NG for that question, though...
TTFN,
Tarkin
no....@no.uce.bellatlantic.net
wrote:
Did that a long time ago using S100, worked pretty good though
as I developed the idea having each share some task loading
via common ram was a performance improvement. The problem
was it was a lot of hardware to do that back in 81-83. with dense
FPGAs with memory it's interesting again. the big difference
between now and then is we have much better compiler technology
to exploit the task parallelism and coroutines between the cpus.
The average PC has not less than 4 maybe many more cpus
(maincpu single maybe dual or quad core, keyboard, Video
accelerator, mouse, and anything semi smart at the other end of the
USB). One key thing is anything that has a net interface or USB
likely contains some amount of embedded CPU.
The idea of multiple cpus oscillates in and out of fashon as cpu
speeds hit bottlenecks. Right now multiple CPUs are back as 3Ghz
seems to be an upper limit.
Allison
>
>TTFN,
> Tarkin
Mr Emmanuel Roche, France
And to get beyond the hacker market segment, you'd
have to find a way to get some graphics on there. Perhaps
some sort of Tektronix video mode, separate from the OS.
Perhaps even pipe it out separate from the cli main screen.
Then, you could do plotting, maybe CAD/CAM, fractals
and such. With even such modest graphical capabilities,
you might capture some research/educational/scientific
marketshare.
Well, Tarkin, I don't know which CP/M computer you used, but it was
not, obviously, an Epson QX-10...
Remember CP/M? Remember the Z-80 CPU? Remember its 4-MHz speed?
Well, between 1986 and 1992 (if I remember correctly), I was
displaying a 3-D maze with MultiFont BASIC under MultiFont CP/M 2.2.
The GDC used was the NEC uPD-7220. At 4-MHz, it was displaying a
picture (640x480) onscreen faster than you could see it drawn...
All the screen drivers for GSX made by Digital Research include code
for this particular chip.
During 15 years, totally alone still using CP/M Plus, I produced
business graphics on a Hewlett-Packard HP-7470A graphics plotter,
after checking them on screen with DR Graph. (And some occasional
drawings with DR Draw.) (Because the plotter had colors, while the
high-resolution driver for the Epson FX-80 were black&White.)
(I also programmed some fractals, using a Logo from Epson that I don't
remember. Epson Mighty Logo?)
Yours Sincerely,
Mr. Emmanuel Roche, France
Tarkin
wrote:
Relax, Mon Ami, most of what you said actually makes my point.
But, the 4 MHz speed is of no concern here, as this thread was
discussing the merits of _accelerated_, Z80 _based_, _hypothetical_
processors. The two Z80 machines I am using now- Apple][ clone
w/ MicroSoft SoftCard Z80 clone, and a Commodore 128- both
noticeably flicker when drawing 80 columns from 'DIR'. So
I wouldn't expect _any_ 4MHz machine to do both OS
and graphics.
But, the GSX and plotter stuff you spoke of are exactly what
inspired my graphics train of thought.
Perhaps you skimmed the thread to quickly?
TTFN,
Tarkin
Tarkin
> On Sat, 29 Nov 2008 09:11:05 -0800 (PST), Tarkin <Tarkin...@gmail.com>
Addendum / correction to something I said earlier:
The OLPC is $199 to give one to child, or
$399 to give one and get one.
http://www.amazon.com/xo/
TTFN,
Tarkin
Tarkin
Correction #2: This an apples and oranges comparison!
Effective speeds for the machines I mentioned are ~2MHz,
nor does the Z80 in either do the actual video generation.
Maybe I was thinking of the Kaypro II before mine died.
TTFN,
Tarkin
Greegor
a faster modern incarnation of the Z-80 cpu, and with
1024x768 graphics, could it be cheap enough
and popular enough to actually compete?
CP/M with graphics enhancements would
never be as good as a cheap pc with linux, would it?
Mr Emmanuel Roche, France
> CP/M with graphics enhancements would
> never be as good as a cheap pc with linux, would it?
Only a Newbie could have written such a stupid sentence.
As I wrote in 2005:
> Should I remind you that you can do only 6 things
> with a computer, be it a Sinclair ZX-80 or a Connection Machine:
> 1) word processing
> 2) programming
> 3) spread-sheet
> 4) database
> 5) communications
> 6) graphics
> As long as those 6 needs are satisfied, the computer is a
> useful tool. And there are/were standards for those 6 needs
> under CP/M: WordStar is obviously the standard for word
> processors; BASIC is obviously the standard for programming,
> MultiPlan is obviously the standard for spread-sheet, dBase II
> is obviously the standard for database; XMODEM is obviously
> the (lowest common denominator) standard for communications;
> and finally GSX is obviously the standard graphics system for CP/M.
Now, find something, anything, that Linux can do better than CP/M!
(By "graphics enhancements", I think that you are thinking "GUIs"...
Should I remember you that GUIs were invented TO LIMIT the users?)
(That's why all the interesting commands are at the bottom of the less
used pull-down menu. I am sure that you did not know that.)
For me, "graphics enhancements" means that, from time to time, one
need a graphics system. In the case of CP/M, Digital Research invented
GSX, which is a PORTABLE graphics system running under CP/M, CP/M-86,
and MS-DOS. Find me a Linux graphics system running under CP/M...
During 15 years, I was totally alone still using CP/M Plus on a Epson
QX-10, running mainly WordStar 4 and Mallard BASIC. All the CP/M User
Groups I knew died one after one, so I finally read the comp.os.cpm
Newsgroup. As soon as TCP/IP will be available for CP/M, I will go
back to CP/M Plus. Probably CP/M-86 Plus, to use cheap IBM Clown
hardware at current speeds. My custom CP/M-86 Plus system runs at 400-
MHz and, each time I use it, I notice how much faster than my Old
Faithful Epson QX-10 it is. I hope to keep using CP/M Plus another 15
years. I have used IBM Mainframes, minicomputers, the Sinclair ZX-80,
various microcomputers, and finally the IBM Clown, first with MeSs-DOS
and now under WinDoze: CP/M Plus is, clearly, without any contest, the
best OS for Single user, and Concurrent CP/M for multi user.
Mr. Emmanuel Roche, France
no....@no.uce.bellatlantic.net
<roch...@laposte.net> wrote:
>Greegor wrote:
>
>> CP/M with graphics enhancements would
>> never be as good as a cheap pc with linux, would it?
>
>Only a Newbie could have written such a stupid sentence.
Besides being rude that was seriously in error.
Linux can provide a richer programming envronment, better IO handling,
and a selection of file systems to store on that are considerably more
robust that the CP/M file system.
>
>As I wrote in 2005:
>
>> Should I remind you that you can do only 6 things
>> with a computer, be it a Sinclair ZX-80 or a Connection Machine:
>> 1) word processing
>> 2) programming
>> 3) spread-sheet
>> 4) database
>> 5) communications
>> 6) graphics
>> As long as those 6 needs are satisfied, the computer is a
>> useful tool. And there are/were standards for those 6 needs
>> under CP/M: WordStar is obviously the standard for word
>> processors; BASIC is obviously the standard for programming,
>> MultiPlan is obviously the standard for spread-sheet, dBase II
>> is obviously the standard for database; XMODEM is obviously
>> the (lowest common denominator) standard for communications;
>> and finally GSX is obviously the standard graphics system for CP/M.
>
>Now, find something, anything, that Linux can do better than CP/M!
Compile a application that he source is near 1gb and the object is
nearly 4Mb compressed. Even the most enhanced CP/M or its kin had
problems with fils greater than 32MB.
Linux is not a GUI, the GUI is actually optional. Myself I prefer the
command line interface as its concise and and not at all complex.
The nicer part of it is be it linux or V6 unix or DEC Ultrix the
command line interface works the save with largely the same
with very similar/same commands.
I enjoy cp/m and was a late adoptor of PC hardware in general
and winders later still. Despite being the trailing endge I was
running the fastest CP/M hardware possible with 8 and 10mhz
parts. Fast CP/M does make doing work genrally a lot more
pleasent and far more productive. While funs or useful I find my
4mhz kaypro ploddingly slow compared to the 8mhz Visual 1050
with a hard disk and monochrome graphics.
I still have a few of the 7220 GDCs laying around, ok chip for it's
time. Certainly was a better part than the crude PC graphics of the
early 80s.
>(By "graphics enhancements", I think that you are thinking "GUIs"...
>Should I remember you that GUIs were invented TO LIMIT the users?)
>(That's why all the interesting commands are at the bottom of the less
>used pull-down menu. I am sure that you did not know that.)
>
>For me, "graphics enhancements" means that, from time to time, one
>need a graphics system. In the case of CP/M, Digital Research invented
>GSX, which is a PORTABLE graphics system running under CP/M, CP/M-86,
>and MS-DOS. Find me a Linux graphics system running under CP/M...
I'd like to see that. The box I use for graphics work is a dual
screen system with each 22" screen running 1680x1024 with full color
depth and the mouse can seamlessly traverse to two screens as if they
were one very wide one. To do that with Z80 would be asking it to
manage and manipulate over 128Mb of graphic memory. Isn't going to
happen.
However if the key word is portable, Xwindows has been run on at least
10 different archectures I know of and likely more. That is portable.
Can it run on a Z80, well yes though its going to be slow. To do the
old 640x480x8bit VGA screen requires a mere 307200 Bytes of video
ram. That in round number 4.7 times more ram than a z80 can address.
Don't forget you need some space for the OS and application.
Paging and other banking schems get around that at the expense of
speed.
>During 15 years, I was totally alone still using CP/M Plus on a Epson
>QX-10, running mainly WordStar 4 and Mallard BASIC. All the CP/M User
>Groups I knew died one after one, so I finally read the comp.os.cpm
>Newsgroup. As soon as TCP/IP will be available for CP/M, I will go
>back to CP/M Plus. Probably CP/M-86 Plus, to use cheap IBM Clown
IP for CP/M has been around for years. Interfaces to ethernet even
existed for S100. The problem is that there is nothing in the CP/M
realm [other than ham radio packet software] that knows how to use
it. Also like much of the CP/M realm and hardware interface for the
IO will be unique or limited ot PPP or SLIP over serial line since
there is some basic Z80 based hardware standards there.
if your running on X86 cpus (PCs) then it's been there for years
and there with hardware support. DOS had networking for years
(DECnet, Lanman, SPX, and IP). Those could be fitted to CP/M+
for x86.
>hardware at current speeds. My custom CP/M-86 Plus system runs at 400-
>MHz and, each time I use it, I notice how much faster than my Old
>Faithful Epson QX-10 it is. I hope to keep using CP/M Plus another 15
>years. I have used IBM Mainframes, minicomputers, the Sinclair ZX-80,
>various microcomputers, and finally the IBM Clown, first with MeSs-DOS
>and now under WinDoze: CP/M Plus is, clearly, without any contest, the
>best OS for Single user, and Concurrent CP/M for multi user.
>
>Mr. Emmanuel Roche, France
Nothing beats fast CPU, lot of ram and a huge amount of disk.
Allison
Howard Harte
> On Sun, 30 Nov 2008 10:59:35 -0800 (PST), "Mr Emmanuel Roche, France"
>
>
> - Show quoted text -
I just uploaded my Vector Graphic Z80 SBC design based on the Xilinx
Spartan-3E Starter kit (US$149) to www.opencores.org:
http://www.opencores.org/projects.cgi/web/vg_z80_sbc/overview
All the sources are there. The project file for ISE is missing at the
moment. I want to re-synthesize and verify everything again on the
S3E Starter Kit, and then I'll post it. I will also post FPGA .BIT
and .MCS files for programming the board. Aside from that, you will
need to program the FLASH on the Spartan 3E Kit with the Vector
Monitor. I'll remember how I did that, and post instructions.
Like I said before, the DDR SDRAM interface is flaky, but the rest
seems to work well. The system runs at 25MHz, but runs somewhat
faster than a real Z80 at 25MHz since the Z80 Core executes
instuctions in fewer cycles than the original.
-Howard
Mr Emmanuel Roche, France
> Linux can provide a richer programming envronment, better IO handling,
> and a selection of file systems to store on that are considerably more
> robust that the CP/M file system.
I like this "richer"... meaning "bloated"! As for "better I/O
handling", it depends what you consider difficult. Finally, who need
"a selection of file systems"? You don't like something small, simple,
and fast? You prefer something big, difficult, and slow? Not me,
despite having started with IBM Mainframes. Maybe a few people need to
drive a truck but, for myself, a bicycle or a small car is enough.
> >Now, find something, anything, that Linux can do better than CP/M!
>
> Compile a application that he source is near 1gb and the object is
> nearly 4Mb compressed. Even the most enhanced CP/M or its kin had
> problems with fils greater than 32MB.
And who needs a source file "near 1 GigaByte"? Who needs a 4 MegaByte
object file? A truck driver, or a Personal Computer user? Review what
I have done and published on the comp.os.cpm Newsgroup during those
last 10 years. Did I ever reach the limit of CP/M?
> > (..). Find me a Linux graphics system running under CP/M...
>
> I'd like to see that. The box I use for graphics work is a dual
> screen system with each 22" screen running 1680x1024 with full color
> depth and the mouse can seamlessly traverse to two screens as if they
> were one very wide one. To do that with Z80 would be asking it to
> manage and manipulate over 128Mb of graphic memory. Isn't going to
> happen.
From everything that you write, your ignorance demonstrates that you
never used the NEC uPC-7220 GDC, only heard about it. For your
information (I am too kind), it manages its own screen memory, which
is totally separate from the RAM of CP/M. You don't seem to remember
that I wrote that, on the Epson QX-10, MultiFont CP/M had a 63K TPA,
along with 640x480 graphics AT THE SAME TIME. I am sure that you never
managed to do that. Finally, the limitations of the NEC uPD-7220 GDC
are 2.000x2.000 pixels... Not too bad, in 1983! So, if you had used it
25 years ago, you could have had your two 22" screens working with a
S-100 Bus system!
But who would need two 22" screens? My main computer has had a 12"
inch screen for the last 10 years. This Windows 98SE computer has a
14" screen, because the seller erroneously told me that it was a 12"
screen. By definition, a screen will ALWAYS be inferior to an eye.
What matter is the output device. Pictures go away, printed matter
stays. During 15 years, I produced business graphics. The 3,000x10,000
pixels of the Hewlett-Packard HP-7470A graphics plotter were more than
enough. And I only used 4 colors: RGB and Black. Who needs more?
> Nothing beats fast CPU, lot of ram and a huge amount of disk.
You are talking only about the hardware. A computer is a combination
of hardware and SOFTWARE. As long as they provide the 6 functions,
they are useful. If the software is as big as a truck, is it as useful
as a bicycle?
no....@no.uce.bellatlantic.net
<roch...@laposte.net> wrote:
>Allison wrote:
>
>> Linux can provide a richer programming envronment, better IO handling,
>> and a selection of file systems to store on that are considerably more
>> robust that the CP/M file system.
>
>I like this "richer"... meaning "bloated"! As for "better I/O
No bloat.
>handling", it depends what you consider difficult. Finally, who need
>"a selection of file systems"? You don't like something small, simple,
Journaling filesystem more robust and retains older versions.
Software has evolved. File systems that can handle errors without
becomming corrupt exist, they can handle versioning.
>and fast? You prefer something big, difficult, and slow? Not me,
>despite having started with IBM Mainframes. Maybe a few people need to
>drive a truck but, for myself, a bicycle or a small car is enough.
Difficult no, slow not on reasonable hardware.
>
>> >Now, find something, anything, that Linux can do better than CP/M!
>>
>> Compile a application that he source is near 1gb and the object is
>> nearly 4Mb compressed. Even the most enhanced CP/M or its kin had
>> problems with fils greater than 32MB.
>
>And who needs a source file "near 1 GigaByte"? Who needs a 4 MegaByte
>object file? A truck driver, or a Personal Computer user? Review what
>I have done and published on the comp.os.cpm Newsgroup during those
>last 10 years. Did I ever reach the limit of CP/M?
A commercial application. The source tree to build it is large.
>
>> > (..). Find me a Linux graphics system running under CP/M...
>>
>> I'd like to see that. The box I use for graphics work is a dual
>> screen system with each 22" screen running 1680x1024 with full color
>> depth and the mouse can seamlessly traverse to two screens as if they
>> were one very wide one. To do that with Z80 would be asking it to
>> manage and manipulate over 128Mb of graphic memory. Isn't going to
>> happen.
>
>From everything that you write, your ignorance demonstrates that you
>never used the NEC uPC-7220 GDC, only heard about it. For your
Yes it have and I've designed with it too. You forget I worked for
the same NEC that sold it and I was there and part of the roll out of
the product.
>information (I am too kind), it manages its own screen memory, which
>is totally separate from the RAM of CP/M. You don't seem to remember
>that I wrote that, on the Epson QX-10, MultiFont CP/M had a 63K TPA,
>along with 640x480 graphics AT THE SAME TIME. I am sure that you never
>managed to do that. Finally, the limitations of the NEC uPD-7220 GDC
>are 2.000x2.000 pixels... Not too bad, in 1983! So, if you had used it
>25 years ago, you could have had your two 22" screens working with a
>S-100 Bus system!
the limitations of the GDC were not that bad but multiple planes
beyond 4 to do good color depth were not there and it was dog slow as
the interface wwhat through the chip.. To you missed that a RMW cycle
on the part was 0.8us so to update a single plane could literally take
several seconds assume no processor overhead more is thats added in.
>But who would need two 22" screens? My main computer has had a 12"
>inch screen for the last 10 years. This Windows 98SE computer has a
>14" screen, because the seller erroneously told me that it was a 12"
>screen. By definition, a screen will ALWAYS be inferior to an eye.
>What matter is the output device. Pictures go away, printed matter
>stays. During 15 years, I produced business graphics. The 3,000x10,000
>pixels of the Hewlett-Packard HP-7470A graphics plotter were more than
>enough. And I only used 4 colors: RGB and Black. Who needs more?
How about someone that is doing builds, working on documentation for
it and all thee thing one needs to do in the real world.
As to more colors... your kidding. Sure I work with terminals in mono
all the time but when I wnat color I want color not just primaries.
>
>> Nothing beats fast CPU, lot of ram and a huge amount of disk.
>
>You are talking only about the hardware. A computer is a combination
>of hardware and SOFTWARE. As long as they provide the 6 functions,
>they are useful. If the software is as big as a truck, is it as useful
>as a bicycle?
Seriously I still use a EPSON PX8 as one of my protable computers.
Your singing to the gallery. I also have a ASUS Eee PC. The latter
can do wireless networking, graphics,full motion video and it runs
Z80emu at the same time. I appreciate what can be done with more
sophisticed software expecially when you need to interact with a group
of engineers on a project.
For standalone projects using sneakernet CP/M is fine.
Allison
no....@no.uce.bellatlantic.net
<roch...@laposte.net> wrote:
>Greegor wrote:
>
>> CP/M with graphics enhancements would
>> never be as good as a cheap pc with linux, would it?
>
>Only a Newbie could have written such a stupid sentence.
>
>As I wrote in 2005:
>
>> Should I remind you that you can do only 6 things
>> with a computer, be it a Sinclair ZX-80 or a Connection Machine:
>> 1) word processing
>> 2) programming
>> 3) spread-sheet
>> 4) database
>> 5) communications
>> 6) graphics
>> As long as those 6 needs are satisfied, the computer is a
>> useful tool. And there are/were standards for those 6 needs
>> under CP/M: WordStar is obviously the standard for word
>> processors; BASIC is obviously the standard for programming,
>> MultiPlan is obviously the standard for spread-sheet, dBase II
>> is obviously the standard for database; XMODEM is obviously
>> the (lowest common denominator) standard for communications;
>> and finally GSX is obviously the standard graphics system for CP/M.
>
>Now, find something, anything, that Linux can do better than CP/M!
>
Here is an example I used to communicate with an amateur in Italy this
AM using low (less than 5W) power from the USA.
Run PSK31 a sound card enabled digital mode for amateur radio. Most
all of the PSK31 programs do the following:
Maintain Keyboard input as it is a text mode communications form so
any key strokes in transmit mode are encoded.
Maintain a real time specturm display aka waterfall during recieve.
Signal strength is encoded as color change.
In recieve mode it's also decoding the spectrum to text, up to 8
channels at the same time and displaying this on the same screen
the waterfall is on.
Perform the needed DSP operations in real time for recieve decode ot
transmit encode.
Control the radio (frequency, power, band select and other parameters)
via serial port.
And do this remotely via an IP connection that includes VIOP if
desired.
It takes an OS that can handle a lot of tasks to do this on fast
hardware. The other way to do this is with multiple cpus one of which
is a fast DSP cpu.
Allison
Mr Emmanuel Roche, France
> Journaling filesystem more robust and retains older versions.
> Software has evolved. File systems that can handle errors without
> becomming corrupt exist, they can handle versioning.
Excuse me, Allison, but this is the comp.os.cpm Newsgroup. If you want
a journaling file system, why don't you make one? Why are you always
criticising, yet never doing anything, in the 10 years that I have
participated to the comp.os.cpm Newsgroup? Half your messages start
with "30 years ago, I made..." (not love, obviously!). You are a
person of the past. Me, I am a person of the present: that's why I am
working (contrary to you) on making TCP/IP run under CP/M.
> A commercial application. The source tree to build it is large.
Ever heard about overlays, Oh you Oldie?
> the limitations of the GDC were not that bad but multiple planes
> beyond 4 to do good color depth were not there and it was dog slow as
> the interface wwhat through the chip.. To you missed that a RMW cycle
> on the part was 0.8us so to update a single plane could literally take
> several seconds assume no processor overhead more is thats added in.
Again, this shows your ignorance: the NEC uPD-7220 GDC could only
manage 8 colors. We are talking of 25 years ago. How many screens
would have displayed "good color depth" 25 years ago? I am afraid that
you are mixing everything, what you did 30 years and the present. Why,
insted of critisising 25 years old hardware, have you not made a S-100
Bus to display VGA graphics with your "good color depth"?
> How about someone that is doing builds, working on documentation for
> it and all thee thing one needs to do in the real world.
>
> As to more colors... your kidding. Sure I work with terminals in mono
> all the time but when I wnat color I want color not just primaries.
Then use Windows or whatever. (You need "good color depth" to "work on
documentation"? "In the real world?" Then, how comes you tolerate
writing this stuff in Black&White, using USASCII characters? Where are
your colors?)
> Seriously I still use a EPSON PX8 as one of my protable computers.
> Your singing to the gallery. I also have a ASUS Eee PC. The latter
> can do wireless networking, graphics,full motion video and it runs
> Z80emu at the same time. I appreciate what can be done with more
> sophisticed software expecially when you need to interact with a group
> of engineers on a project.
If you "need to interact with a group of engineers on a project", use
a network. Guess why Usenet was created? Why are you using it? In
Black&White?
Mr Emmanuel Roche, France
> Here is an example I used to communicate with an amateur in Italy this
> AM using low (less than 5W) power from the USA.
>
> Run PSK31 a sound card enabled digital mode for amateur radio. Most
> all of the PSK31 programs do the following:
Haaa... This stubborn bad faith!
Excuse me, Allison, but (from your description) I have problem
understanding if you are listing the technical abilities of a card, or
of a program. After re-reading your stuff, I think that you are
quoting the abilities of a card.
But, then, Allison, what are you waiting to put such a chip on a CP/M
S-100 Bus card?
> It takes an OS that can handle a lot of tasks to do this on fast
> hardware. The other way to do this is with multiple cpus one of which
> is a fast DSP cpu.
And Concurrent CP/M is not CP/M? What a stubborn bad faith!
(According to the doc, MP/M-II switch tasks in 600 to 900 microseconds
(on a 4-MHz CPU). If you were using MP/M-86 or Concurrent CP/M, you
could use current speed CPUs. Me, I have a 400-MHz CPU, and I am happy
with it. And, anyway, radio HAM is quite a specialized area, since it
was never mentioned during 10 years in the comp.os.cpm Newsgroup.
Else, you could use a galena radio...)
Glen Herrmannsfeldt
(snip)
> I just uploaded my Vector Graphic Z80 SBC design based on the Xilinx
> Spartan-3E Starter kit (US$149) to www.opencores.org:
> http://www.opencores.org/projects.cgi/web/vg_z80_sbc/overview
> All the sources are there. The project file for ISE is missing at the
> moment.
(snip)
Which size FPGA do you have? (I believe more than one size is available.)
> Like I said before, the DDR SDRAM interface is flaky, but the rest
> seems to work well. The system runs at 25MHz, but runs somewhat
> faster than a real Z80 at 25MHz since the Z80 Core executes
> instuctions in fewer cycles than the original.
It looks like the older boards used SRAM, but the 3E uses DDR, though
much bigger. (I believe 16MB. Should be big enough for any CP/M system.)
Also, the 3E has ethernet. One could implement NFS for CP/M.
-- glen
Glen Herrmannsfeldt
(snip)
> IP for CP/M has been around for years. Interfaces to ethernet even
> existed for S100. The problem is that there is nothing in the CP/M
> realm [other than ham radio packet software] that knows how to use
> it. Also like much of the CP/M realm and hardware interface for the
> IO will be unique or limited ot PPP or SLIP over serial line since
> there is some basic Z80 based hardware standards there.
I would like to see NFS for CP/M. For one, it is (or can be) UDP
based, much easier than TCP. Also, it avoids much of the problem
with small slow disk drives. I believe it avoids many of the
limitations that CP/M file systems would otherwise have. NFS clients
don't know about sectors and such, but request specific blocks of
the file. It should be possible to put a very large disk on
an NFS server and access it through CP/M.
-- glen
Axel Berger
>for myself, a bicycle or a small car is enough.
I suppose that's why you've got a big Mercedes estate with big towing
capability.
Mr Emmanuel Roche, France
> I suppose that's why you've got a big Mercedes estate with big towing
> capability.
Hahaha! Congratulations for the discretion!
Yes, I bought a 15-years old Mercedes 300TE (estate, the luxury
version with leather seats and "tempomat", fitted to only 20% of
Mercedes cars) just one month before the 3rd Oil Crisis...
Since then, it has stayed in the garage, along with my 36-years old
city car, a Audi 50. (The predecessor of the VolksWagen Golf or
Rabbit.)
Following my knocking over by a car, I was ordered by a Doctor to make
more physical exercise. So, 3 times per week, I go to the next city by
foot (1 hour 15 minutes), rest a little in a cybercafe, then go back
home.
The other days of the week, I use my moped (a Peugeot Fox) for
shopping or moving. I spend less than 5 Euros of petrol per week.
For the last 40 years, I have been going to England almost every year.
This Mercedes is the perfect vehicle for those kinds of long-range
driving, especially with its "tempomat" (speed regulator) which
changes completely the way you drive. It also has an automatic gear,
which, in France, is reserved for luxury vehicles. (This car was used
by one big French company to transport its biggest customers to/from
airports: that's why they took an estate, for the luggages.)
This is the first car that I buy (I inherited my grand-father's Audi).
Judging by the way people were looking at me at the petrol stations,
it seems to be good looking.
Howard Harte
wrote:
The Spartan 3E kit has an XC3S500E on it, which is 500K gates. The
CPU core is the biggest part, and the whole design uses up about 70%
of the FPGA.
-Howard
Glen Herrmannsfeldt
(snip)
> The Spartan 3E kit has an XC3S500E on it, which is 500K gates. The
> CPU core is the biggest part, and the whole design uses up about 70%
> of the FPGA.
The NEXYS2 board also has a Spartan3E on it, available in either
500K gate or 1200K gate versions.
So far, it seems that the SPARTAN3E board only comes with the 300K version.
It does have ethernet, which NEXYS2 doesn't have. Also, SPARTAN3E has an
LCD display, where NEXYS2 has an LED display.
I would expect an 8080 core to be much smaller than a Z80 core.
-- glen
Peter Dassow
> Allison wrote:
>
>> Linux can provide a richer programming envronment, better IO handling,
>> and a selection of file systems to store on that are considerably more
>> robust that the CP/M file system.
>
> I like this "richer"... meaning "bloated"! As for "better I/O
> handling", it depends what you consider difficult. Finally, who need
> "a selection of file systems"? You don't like something small, simple,
> and fast? You prefer something big, difficult, and slow? Not me,
> despite having started with IBM Mainframes. Maybe a few people need to
> drive a truck but, for myself, a bicycle or a small car is enough.
>
Why so polarizing ?
Please calm down, this is for almost all here just a hobby.
This is not a discussion which ends in a good result.
I agree with you, that CP/M - related to the resources it uses - is
still good enough to work with. But. If you like to get more
functionality, more comfort, and something like WYSIWYG which would be
really difficult with CP/M, even with GSX-80 (and, beside the Amstrad
8512 and the Epson QX-10, what CP/M machine owns a working graphics card
GSX ?), Linux or Windows would a better choice.
So both of you are right, depending on the point of view.
I really like your contributions of old articles, but I dislike such
comments because it is a bit destructive.
Regards to france
Peter
no....@no.uce.bellatlantic.net
<roch...@laposte.net> wrote:
>What I like about Allison J. Parent is its stubborn bad faith.
>
>> Journaling filesystem more robust and retains older versions.
>> Software has evolved. File systems that can handle errors without
>> becomming corrupt exist, they can handle versioning.
>
>Excuse me, Allison, but this is the comp.os.cpm Newsgroup. If you want
>a journaling file system, why don't you make one? Why are you always
>criticising, yet never doing anything, in the 10 years that I have
>participated to the comp.os.cpm Newsgroup? Half your messages start
>with "30 years ago, I made..." (not love, obviously!). You are a
>person of the past. Me, I am a person of the present: that's why I am
>working (contrary to you) on making TCP/IP run under CP/M.
Well I do make things and dont advertizer here. As to the other
well, I did do it 30 years ago and it's still valid now. I presume I
have to do it yet again for it to mean something to you? I'm still
running it and it still works very well so why make another.
>
>> A commercial application. The source tree to build it is large.
>
>Ever heard about overlays, Oh you Oldie?
Yep, but the application doesn't have a disk.
>
>> the limitations of the GDC were not that bad but multiple planes
>> beyond 4 to do good color depth were not there and it was dog slow as
>> the interface wwhat through the chip.. To you missed that a RMW cycle
>> on the part was 0.8us so to update a single plane could literally take
>> several seconds assume no processor overhead more is thats added in.
>
>Again, this shows your ignorance: the NEC uPD-7220 GDC could only
>manage 8 colors. We are talking of 25 years ago. How many screens
>would have displayed "good color depth" 25 years ago? I am afraid that
You would be suprized how many good color monitors there were. They
werent cheap but they existed and were not hard to find.
>you are mixing everything, what you did 30 years and the present. Why,
>insted of critisising 25 years old hardware, have you not made a S-100
>Bus to display VGA graphics with your "good color depth"?
Yes. I interfaced a VGA board directly to S100, worked well too.
>> How about someone that is doing builds, working on documentation for
>> it and all thee thing one needs to do in the real world.
>>
>> As to more colors... your kidding. Sure I work with terminals in mono
>> all the time but when I wnat color I want color not just primaries.
>
>Then use Windows or whatever. (You need "good color depth" to "work on
>documentation"? "In the real world?" Then, how comes you tolerate
Ick winders! Don't know about you but my documentation has real
imaged and graphics embedded in them.
>writing this stuff in Black&White, using USASCII characters? Where are
>your colors?)
In my words of course.
>> Seriously I still use a EPSON PX8 as one of my protable computers.
>> Your singing to the gallery. I also have a ASUS Eee PC. The latter
>> can do wireless networking, graphics,full motion video and it runs
>> Z80emu at the same time. I appreciate what can be done with more
>> sophisticed software expecially when you need to interact with a group
>> of engineers on a project.
>
>If you "need to interact with a group of engineers on a project", use
>a network. Guess why Usenet was created? Why are you using it? In
>Black&White?
Funny thing you missed the whole point. I do work with engineers on
no less than two continents. Networking is wonderful especially when
it includes images.
Allison
no....@no.uce.bellatlantic.net
wrote:
I use a Visual 1050 (CPM+) machine and that has for CP/M decent res
monochrome grpahics and part of the tool kit for that is a full suite
of the DRI graphic products and OS tools. Fun to work with.
My other CP/M machine has a VGA card interfaced to the S100 bus and
it's pretty cool to see CCP in orange text and applications in green.
Even for text color is interesting to play with.
Allison
>Regards to france
> Peter
no....@no.uce.bellatlantic.net
I was looking at the various IP stacks to do that. The interface to
the bios is as you say easy, it's all blocks. The problem is the host
side. It you do NFS then you have to implment everything above IP
as the assumed file system is usually linux based. The shortcut is
a host program to implment container files that can be formatted
as if they were CP/M on disk format. One caveat is the container
files have to be fixed size as CP/M doesn't have a easy way to deal
with dynamicaly changing disk sizes. OF copurse doing it over a
serial line is nearly the same thing and saves the network stack.
Allison
>-- glen
no....@no.uce.bellatlantic.net
<roch...@laposte.net> wrote:
>Allison "30 years ago..." Parent wrote:
>
>> Here is an example I used to communicate with an amateur in Italy this
>> AM using low (less than 5W) power from the USA.
>>
>> Run PSK31 a sound card enabled digital mode for amateur radio. Most
>> all of the PSK31 programs do the following:
>
>Haaa... This stubborn bad faith!
A clear example of what you have no understanding of.
I'll help you its software as in Software Defined Radio, do a Google
on "SDR". You may find it interesting how computers and radio
are very related.
>
>Excuse me, Allison, but (from your description) I have problem
>understanding if you are listing the technical abilities of a card, or
>of a program. After re-reading your stuff, I think that you are
>quoting the abilities of a card.
It's a use of common PCs plus some sophisticated software.
The result is that the signal processing of the computer can reasult
in capability that exceeds the radio plus human ears.
>But, then, Allison, what are you waiting to put such a chip on a CP/M
>S-100 Bus card?
If you understood what your talking about you'd realize the folly of
those words.
>> It takes an OS that can handle a lot of tasks to do this on fast
>> hardware. The other way to do this is with multiple cpus one of which
>> is a fast DSP cpu.
>
>And Concurrent CP/M is not CP/M? What a stubborn bad faith!
It isn't. Different beast inside.
>
>(According to the doc, MP/M-II switch tasks in 600 to 900 microseconds
Not fast enough. Why slow Z80. In 1982 you could get 6 and 8mhz
Z80s.
>(on a 4-MHz CPU). If you were using MP/M-86 or Concurrent CP/M, you
>could use current speed CPUs. Me, I have a 400-MHz CPU, and I am happy
>with it
Wow 400mhz. I just retired two of them as too slow. Those are what
we call recyle fodder here. However I do keep a pair of 486/66 as
those have the needed bus interfaces to handle a wide variety of old
cards to do things like read floppies and old hard disks. When doing
stuff like that too fast can be a problem.
Current speed CPUs might still run Concurrent if you had the needed
drivers to enable the busses for the various devices. Otherwise it's
way to crippled. You need a current OS to support those devices.
>. And, anyway, radio HAM is quite a specialized area, since it
>was never mentioned during 10 years in the comp.os.cpm Newsgroup.
Really I'll bet it's come up at least a few times in thelast year.
Remember the topic covering KA9Q IP? Yes, hams use computers
and in recent years computers are very much part of the radio to the
point that most contain DSP and multiple MPUs plus interface to the
home computer to do more things. One radio I designed is actually
useless without a computer to do the processing to translate RF
something that might come out of the speaker. Again see SDR.
>Else, you could use a galena radio...)
Still needs a BFO to recieve SSB signals.
Allison
Steven Hirsch
> The Spartan 3E kit has an XC3S500E on it, which is 500K gates. The
> CPU core is the biggest part, and the whole design uses up about 70%
> of the FPGA.
Drat. I assume it will not fit on my 200K gate Spartan 3, then.
Glen Herrmannsfeldt
(Snip, I wrote)
>>I would like to see NFS for CP/M. For one, it is (or can be) UDP
>>based, much easier than TCP. Also, it avoids much of the problem
>>with small slow disk drives. I believe it avoids many of the
>>limitations that CP/M file systems would otherwise have. NFS clients
>>don't know about sectors and such, but request specific blocks of
>>the file. It should be possible to put a very large disk on
>>an NFS server and access it through CP/M.
> I was looking at the various IP stacks to do that. The interface to
> the bios is as you say easy, it's all blocks. The problem is the host
> side. It you do NFS then you have to implment everything above IP
> as the assumed file system is usually linux based. The shortcut is
> a host program to implment container files that can be formatted
> as if they were CP/M on disk format. One caveat is the container
> files have to be fixed size as CP/M doesn't have a easy way to deal
> with dynamicaly changing disk sizes. OF copurse doing it over a
> serial line is nearly the same thing and saves the network stack.
I was assuming that the server would be unix-like. NFS is usually
a little higher in the file processing chain than other file systems,
with interesting results.
I remember with HP-UX many years ago, the native file systems had a
limit of something like 17 characters for file names, but for NFS
it was the server limit that applied. I believe it will avoid
some of the limits that CP/M file systems traditionally have.
File size limits might still be there, but the disk size
limit should be gone. I would have to look through the
BIOS some more to be sure.
I believe NFS allows the client to specify the block size.
NFS usually likes large blocks, like 4K or 8K, but I believe
that it can run at 128. Much of the complications of the CP/M
bios used for local disk files would not be needed, which might
just about make up for the extra needed for UDP/IP.
Diskless booting of CP/M should be easier than diskless
booting of unix!
-- glen
no....@no.uce.bellatlantic.net
Save for all the applicatiopns adhere with the 8.3 and also the Bdos
enforces it. The bios only implments the hardware interface to the
BDOS so a NFS would still be limited to the bdos and bios limitations.
>
>File size limits might still be there, but the disk size
>limit should be gone. I would have to look through the
>BIOS some more to be sure.
Definate no. the bios has to know the device size check teh entries
into DPH and DPB, though they can be overlaid to appear dynamic.
>
>I believe NFS allows the client to specify the block size.
>NFS usually likes large blocks, like 4K or 8K, but I believe
>that it can run at 128. Much of the complications of the CP/M
>bios used for local disk files would not be needed, which might
>just about make up for the extra needed for UDP/IP.
Leave the block size for the CP/M side as deblocking. CP/M
can handle deblocking any size "sector".
>
>Diskless booting of CP/M should be easier than diskless
>booting of unix!
Dead simple. It's a mater of loading an image of CP/M and
jumping to the start address. I still do that for testing new CP/M
systems, a serial downloader that puts bytes in ram and the right
(selected) place and then when done jumps to teh start address
and it's up. Really not much different from copying it from a rom
to ram or the system tracks to ram.
Allison
>-- glen
Bi...@sunsouthwest.com
<roch...@laposte.net> wrote:
>Now, find something, anything, that Linux can do better than CP/M!
play (not to mention record and/or edit) full motion video
network and internet (browsing, file transfer, etc)
did you ever see a scanner setup for CP/M? Scanning is nice.
take advantage of 'recent' Intel cpu's - say, since the 386
CP/M was a rudimentary system long ago left in the dust
To students of history its story may be fascinating but as a
functional product, it finds fewer and fewer rational uses.
Bill
Glen Herrmannsfeldt
(snip)
>>I was assuming that the server would be unix-like. NFS is usually
>>a little higher in the file processing chain than other file systems,
>>with interesting results.
>>I remember with HP-UX many years ago, the native file systems had a
>>limit of something like 17 characters for file names, but for NFS
>>it was the server limit that applied. I believe it will avoid
>>some of the limits that CP/M file systems traditionally have.
> Save for all the applicatiopns adhere with the 8.3 and also the Bdos
> enforces it. The bios only implments the hardware interface to the
> BDOS so a NFS would still be limited to the bdos and bios limitations.
Yes, it would have to go at a higher level for other filename
lengths to work.
>>File size limits might still be there, but the disk size
>>limit should be gone. I would have to look through the
>>BIOS some more to be sure.
> Definate no. the bios has to know the device size check teh entries
> into DPH and DPB, though they can be overlaid to appear dynamic.
For NFS, I believe the only reason to know the device size
is to report the available space. (Which on some systems
is negative, as they use too few bits to do the calculation.)
>>I believe NFS allows the client to specify the block size.
>>NFS usually likes large blocks, like 4K or 8K, but I believe
>>that it can run at 128. Much of the complications of the CP/M
>>bios used for local disk files would not be needed, which might
>>just about make up for the extra needed for UDP/IP.
> Leave the block size for the CP/M side as deblocking. CP/M
> can handle deblocking any size "sector".
Yes, but allocating 8K for the buffer is a big part of
the available memory.
>>Diskless booting of CP/M should be easier than diskless
>>booting of unix!
> Dead simple. It's a mater of loading an image of CP/M and
> jumping to the start address. I still do that for testing new CP/M
> systems, a serial downloader that puts bytes in ram and the right
> (selected) place and then when done jumps to teh start address
> and it's up. Really not much different from copying it from a rom
> to ram or the system tracks to ram.
For Sun it is done first through TFTP. (The first, and I
believe second, level boot loader is through TFTP. Then
it does an NFS mount to load the actual system.)
I suppose one could do everything through TFTP, though.
Traditionally, unix allows programs to read directories
just like files. NFS doesn't allow that. (It would not
be portable, as the server might have a different directory
format.) I will get out my CP/M book again and think about
how it would work.
-- glen
Mr Emmanuel Roche, France
> >Now, find something, anything, that Linux can do better than CP/M!
>
> play (not to mention record and/or edit) full motion video
You are, obviously, a Newbie.
I don't have a collection of BYTE magazines, but I am 100% sure that
there were advertisements of S-100 Bus boards able to display American
TV films, 25 years ago.
> network and internet (browsing, file transfer, etc)
Then, could you tell me why Digital Reserch made CP/NET?
> did you ever see a scanner setup for CP/M? Scanning is nice.
Of course I saw one. I even used one, 20 years ago! I even remember
some models that fitted on the printhead of a printer, to do the
scanning.
> take advantage of 'recent' Intel cpu's - say, since the 386
Here, I understand less. For me, the 386 is a 32-bit chip. CP/M was 8
bits, CP/M-86 was 16 bits. As far as I know, all of them are able to
display characters. So, what would 32-bits CPU add? (Even 30 years
ago, 32-bits computers existed. Ever heard of IBM? Yet, CP/M was so
much better, simpler, smaller, faster, cheaper than the monsters of
the day that it became the standard Operating System of
microcomputers.)
> CP/M was a rudimentary system long ago left in the dust
Really? Then, what are you doing in the comp.os.cpm Newsgroup, in
December 2008? Waiting for Santa Claus?
> To students of history its story may be fascinating but as a
> functional product, it finds fewer and fewer rational uses.
Only a Newbie could have written such a stupid sentence.
As I wrote in 2005:
> Should I remind you that you can do only 6 things
> with a computer, be it a Sinclair ZX-80 or a Connection Machine:
> 1) word processing
> 2) programming
> 3) spread-sheet
> 4) database
> 5) communications
> 6) graphics
> As long as those 6 needs are satisfied, the computer is a
> useful tool. And there are/were standards for those 6 needs
> under CP/M: WordStar is obviously the standard for word
> processors; BASIC is obviously the standard for programming,
> MultiPlan is obviously the standard for spread-sheet, dBase II
> is obviously the standard for database; XMODEM is obviously
> the (lowest common denominator) standard for communications;
> and finally GSX is obviously the standard graphics system for CP/M.
Now, find something, anything, that Linux can do better than CP/M!
Yours Sincerely,
Mr. Emmanuel Roche, France
no....@no.uce.bellatlantic.net
<g...@ugcs.caltech.edu> wrote:
>no....@no.uce.bellatlantic.net wrote:
>(snip)
>
>>>I was assuming that the server would be unix-like. NFS is usually
>>>a little higher in the file processing chain than other file systems,
>>>with interesting results.
>
>>>I remember with HP-UX many years ago, the native file systems had a
>>>limit of something like 17 characters for file names, but for NFS
>>>it was the server limit that applied. I believe it will avoid
>>>some of the limits that CP/M file systems traditionally have.
>
>> Save for all the applicatiopns adhere with the 8.3 and also the Bdos
>> enforces it. The bios only implments the hardware interface to the
>> BDOS so a NFS would still be limited to the bdos and bios limitations.
>
>Yes, it would have to go at a higher level for other filename
>lengths to work.
>
>>>File size limits might still be there, but the disk size
>>>limit should be gone. I would have to look through the
>>>BIOS some more to be sure.
>
>> Definate no. the bios has to know the device size check teh entries
>> into DPH and DPB, though they can be overlaid to appear dynamic.
>
>For NFS, I believe the only reason to know the device size
>is to report the available space. (Which on some systems
>is negative, as they use too few bits to do the calculation.)
I was thinking more about the CP/M side as that is the side with the
more specific limitations. The DPH and DPB define what the disk
looks like such as directory size, check buffer size and over all disk
size plus things like allocation size.
>>>I believe NFS allows the client to specify the block size.
>>>NFS usually likes large blocks, like 4K or 8K, but I believe
>>>that it can run at 128. Much of the complications of the CP/M
>>>bios used for local disk files would not be needed, which might
>>>just about make up for the extra needed for UDP/IP.
>
>> Leave the block size for the CP/M side as deblocking. CP/M
>> can handle deblocking any size "sector".
>
>Yes, but allocating 8K for the buffer is a big part of
>the available memory.
Always a problem in 64k, one solution is memory banking or paging.
>
>>>Diskless booting of CP/M should be easier than diskless
>>>booting of unix!
>
>> Dead simple. It's a mater of loading an image of CP/M and
>> jumping to the start address. I still do that for testing new CP/M
>> systems, a serial downloader that puts bytes in ram and the right
>> (selected) place and then when done jumps to teh start address
>> and it's up. Really not much different from copying it from a rom
>> to ram or the system tracks to ram.
>
>For Sun it is done first through TFTP. (The first, and I
>believe second, level boot loader is through TFTP. Then
>it does an NFS mount to load the actual system.)
TTFTP is also used on OSx Mac(apple), linux and many more.
I use it to boot a blackfin.
>I suppose one could do everything through TFTP, though.
The T stands for trivial and also no so robust.
>Traditionally, unix allows programs to read directories
>just like files. NFS doesn't allow that. (It would not
>be portable, as the server might have a different directory
>format.) I will get out my CP/M book again and think about
>how it would work.
True. the problem is the directory and file structure is unusable to
CP/M. NFS is likely the most complex for the CP/M side as it would
have to learn a new and more complex filesystem even though it's
simplest for the host. THe other way is a host program that defines
a files of say 8mb as a container and the data in the files is in a
CP/M file structure. The Host program would use IP as a transport
and make the file(s) available in a more CP/M friendly form.
Allison
>
>-- glen
Michael Haardt
get a BDOS extension that traps all calls dealing with a network drive.
If I got things right, you would better run a CP/NET Server on Unix
instead to save the implementation trouble and use UDP as transport.
A minimal UDP implementation including ARP should only be a few pages
of code and CP/NET is already there.
If you simply want remote storage, then I suggest to use a storage
access protocol, like NBD which originated in Linux, iSCSI or something
like that.
Microcomputers work most efficient if intelligent peripheral chips are
used, and networking makes no exception here. A Z80 running CP/M would
profit a lot from offloading the network stack to a network controller
and such things exist already.
Michael
no....@no.uce.bellatlantic.net
<roch...@laposte.net> wrote:
>Bill wrote:
>
>> >Now, find something, anything, that Linux can do better than CP/M!
>>
>> play (not to mention record and/or edit) full motion video
>
>You are, obviously, a Newbie.
>
>I don't have a collection of BYTE magazines, but I am 100% sure that
>there were advertisements of S-100 Bus boards able to display American
>TV films, 25 years ago.
At 30 frames per second and better than 1Mb per frame you have
exceeded the data rate for S100 bus by a factor of 5 or more.
As to the rest of the capability editing full motion video is really
not a posibility with CP/M or for that fact even editing a still
frame.
>
>> network and internet (browsing, file transfer, etc)
>
>Then, could you tell me why Digital Reserch made CP/NET?
It still cannot browse the internet and your still longing for a IP
interface to even connect to the net.
>
>> did you ever see a scanner setup for CP/M? Scanning is nice.
>
>Of course I saw one. I even used one, 20 years ago! I even remember
>some models that fitted on the printhead of a printer, to do the
>scanning.
Not used or useful under CP/M as you can expect about a megabyte
per page at 300 DPI and most CP/M system without a hard disk did not
have disks large enough to hold that much. Those scanning rigs were
sold for and used with DOS on PCs.. I still have one that fits in
place of a Epson MX80 ribbon cartridge. Worked ok for black and white
at low resolution and about 2-3minutes per page. Also the very
definition of slow and barely useful.
>> take advantage of 'recent' Intel cpu's - say, since the 386
>
>Here, I understand less. For me, the 386 is a 32-bit chip. CP/M was 8
>bits, CP/M-86 was 16 bits. As far as I know, all of them are able to
>display characters. So, what would 32-bits CPU add? (Even 30 years
Yes, you really do understand less. speed for one, memory,
application( interprocess) protection. The latter is since so that a
errant program doesnt crash the whole show (CP/M and DOS share the
inability to employ that). And the show stopper is the ability to
have larger than 640K(1mb cor CP/M86) memory spaces to take
advantage of larger memeory arrays and do calulations to higher
precision at far greater speeds.
>ago, 32-bits computers existed. Ever heard of IBM? Yet, CP/M was so
DEC VAX 1978 for the 11/780 was also 32bit. It's big thing was then an
still the ability to protect once users program from the effects of
another while dealing with larger memory arrays. In a multitasking,
multiuser envronment this is a significant requirement.
>much better, simpler, smaller, faster, cheaper than the monsters of
>the day that it became the standard Operating System of
>microcomputers.)
Not better, but it did fit the 64k and floppy envronment well. Of
course DEC had RT11 wich fit the 16bit PDP11 with 32kW of ram
and floppies very well. The point being small computers need small
OSs and CP/M is a small file system and barely an OS by most
standards.
>
>> CP/M was a rudimentary system long ago left in the dust
>
>Really? Then, what are you doing in the comp.os.cpm Newsgroup, in
>December 2008? Waiting for Santa Claus?
>
>> To students of history its story may be fascinating but as a
>> functional product, it finds fewer and fewer rational uses.
>
>Only a Newbie could have written such a stupid sentence.
>
>As I wrote in 2005:
>
>> Should I remind you that you can do only 6 things
>> with a computer, be it a Sinclair ZX-80 or a Connection Machine:
>> 1) word processing
>> 2) programming
>> 3) spread-sheet
>> 4) database
>> 5) communications
>> 6) graphics
>> As long as those 6 needs are satisfied, the computer is a
>> useful tool. And there are/were standards for those 6 needs
>> under CP/M: WordStar is obviously the standard for word
>> processors; BASIC is obviously the standard for programming,
>> MultiPlan is obviously the standard for spread-sheet, dBase II
>> is obviously the standard for database; XMODEM is obviously
>> the (lowest common denominator) standard for communications;
You can communicate with mose code but sending pictures with it is
not easily accomplised. It's the lowest common denominator for radio.
I'm a fan of CP/M and still use it too. However, everything about it
is the lowest common denominator.
>> and finally GSX is obviously the standard graphics system for CP/M.
IT may be but a poll of how many systems could:
Run it
Display it
or take graphic input
You will find while it was a standard it was also relatively unknown.
>Now, find something, anything, that Linux can do better than CP/M!
Still bad faith.
Allison
no....@no.uce.bellatlantic.net
>To use NFS in a useful way, and justify the implementation, CP/M had to
>get a BDOS extension that traps all calls dealing with a network drive.
>If I got things right, you would better run a CP/NET Server on Unix
>instead to save the implementation trouble and use UDP as transport.
>A minimal UDP implementation including ARP should only be a few pages
>of code and CP/NET is already there.
That would be a workable scheme.
>If you simply want remote storage, then I suggest to use a storage
>access protocol, like NBD which originated in Linux, iSCSI or something
>like that.
>
>Microcomputers work most efficient if intelligent peripheral chips are
>used, and networking makes no exception here. A Z80 running CP/M would
>profit a lot from offloading the network stack to a network controller
>and such things exist already.
It would be the most practical way to build it. There are chips like
the PIC18 that might have enough to do that.
Allison
>
>Michael
Bi...@sunsouthwest.com
<roch...@laposte.net> wrote:
>Bill wrote:
>
>> CP/M was a rudimentary system long ago left in the dust
>
>Really? Then, what are you doing in the comp.os.cpm Newsgroup, in
>December 2008? Waiting for Santa Claus?
I have a garage full of it! I wanted to get rid of most of it five
years ago, but my son, an Anthropology professor, thinks
I oughta make some effort to put in writing the things I've
learned from actually using these things as well as behind
the scenes 'what-really-happened' stuff never published
I've mentioned a few things here - more detail about how/why
ms used OS/2 to screw IBM to the floor; how ms really got nerworking
into their products; what happened in that bar in Monteray that took
Kildall from us - maybe I shoulda been a detective or maybe a lawyer.
If you check copyrights, you'll find I built Voice Stress Analyzers
thirty years ago - that was my big project during the MITS years.
(they work, but lying has become epidemic so who cares?)
>> To students of history its story may be fascinating but as a
>> functional product, it finds fewer and fewer rational uses.
>
>Only a Newbie could have written such a stupid sentence.
Somebody here is stupid all right
>As I wrote in 2005:
>
>> Should I remind you that you can do only 6 things
>> with a computer, be it a Sinclair ZX-80 or a Connection Machine:
>> 1) word processing
>> 2) programming
>> 3) spread-sheet
>> 4) database
>> 5) communications
>> 6) graphics
VIDEO (as in multimedia) is a little bit more than GRAPHICS
>Now, find something, anything, that Linux can do better than CP/M!
You are aware most of the set-top DV recorders are running Linux?
Or, maybe you just didn't know
Bill
Bi...@sunsouthwest.com
>Microcomputers work most efficient if intelligent peripheral chips are
>used, and networking makes no exception here. A Z80 running CP/M would
>profit a lot from offloading the network stack to a network controller
>and such things exist already.
At some point, the Z80 ceases to be your CPU and becomes instead
merely a PPU - peripheral processing unit.
I used to use Columbia Data Products SCSI controllers in ISA slots
that used a Z80 as the processor. I've still got what I call the
'worlds fastest 486' (stock) that's running one. Of around two
thousand machines tested and reporting results on Toms Hardware,
this scored #1 without even trying to push it. Surprised me the speed
demons had never figured it out. (SK 096)
In fact I keep it running as evidence - this machine is faster than
any Pentium under 90-100Mhz. That means all that hype about
60s, 66s, and 75Mhz Pentiums was nothing less than a massive
consumer FRAUD. Not that anybody ever cared.
Bill
Michael Haardt
> On 03 Dec 2008 13:10:23 GMT, mic...@moria.de (Michael Haardt) wrote:
>
> >Microcomputers work most efficient if intelligent peripheral chips are
> >used, and networking makes no exception here. A Z80 running CP/M would
> >profit a lot from offloading the network stack to a network controller
> >and such things exist already.
>
> At some point, the Z80 ceases to be your CPU and becomes instead
> merely a PPU - peripheral processing unit.
Well, it does what its best at. A Z80 can't interface a HD floppy
directly very well, so it uses a FDC and a DMA controller. Steal cycles
from its RAM for video, and performance is bad, so you use a video
controller with private RAM. The SIO has a small FIFO buffer to reduce
latency requirements, and so on. In the eighties, all of that was already
common technology, and had people had ethernet back then, they had built
network controllers.
The Z80 family includes many peripheral chips for a reason. These days,
PICs and the like allow to build you own peripheral controllers at low
cost and using low power.
Michael
Mr Emmanuel Roche, France
> I have a garage full of it!
Then, I encourage you to do something with it.
> (...) maybe I shoulda been a detective or maybe a lawyer.
Well, if you never explain, exactly, what you know and how you learned
it, you will remain what you are.
> (...) but lying has become epidemic so who cares?)
As a European, I enjoyed this sentence, after the stubborn lying of
George "War! War!! War!!!" Bush during 8 years... (I am a child of
WWII, so grew up watching "western" films featuring John Wayne. I
would have never imagined to see the USA so hated everywhere in the
world because of one single US President. As I wrote several years
ago, "George Bush will end up in the garbage bin of history". I tried
to find how many years ago I wrote this, but could not find it (it was
in answer to Chuck Falconer).)
> Somebody here is stupid all right
Ok. Let us try to stay polite.
> VIDEO (as in multimedia) is a little bit more than GRAPHICS
Yes, but what is the relation between computer and video? Are you
obliged to use a computer to do everything? Personally, I don't think
so. I still hand write, from time to time.
> You are aware most of the set-top DV recorders are running Linux?
No.
> Or, maybe you just didn't know
Yes, since I have no TeleVision, and no recorders of any kind.
As I wrote, images just go away, while writings stay.
To know what is happening in the world, I read news magazines and hear
the radio. It must have been 25 years since I last lived in a house
with a TV.
I have no "portable phone", either.
I am not totally against progress, but TV and "cell phones" do not
seem, to me, in my humble opinion, good examples of progress.
Mr Emmanuel Roche, France
> As to the rest of the capability editing full motion video is really
> not a posibility with CP/M or for that fact even editing a still
> frame.
But why would you want to "edit full motion video"?
And is "Bill" your first name? I answered Bill, not you. How do you
call that, in American, answering at the place of someone else? Is it
something that good girls do?
> It still cannot browse the internet and your still longing for a IP
> interface to even connect to the net.
It all depends what you call "browse the Internet" and, anyway, you
have done nothing to bring TCP/IP under CP/M.
> Not used or useful under CP/M as you can expect about a megabyte
> per page at 300 DPI and most CP/M system without a hard disk did not
> have disks large enough to hold that much. Those scanning rigs were
> sold for and used with DOS on PCs.. I still have one that fits in
> place of a Epson MX80 ribbon cartridge. Worked ok for black and white
> at low resolution and about 2-3minutes per page. Also the very
> definition of slow and barely useful.
You just forgot that scanners of the CP/M area did not made 300 DPI.
Anyway, I have never thought that they were useful.
> Yes, you really do understand less. speed for one, memory,
> application( interprocess) protection. The latter is since so that a
> errant program doesnt crash the whole show (CP/M and DOS share the
> inability to employ that). And the show stopper is the ability to
> have larger than 640K(1mb cor CP/M86) memory spaces to take
> advantage of larger memeory arrays and do calulations to higher
> precision at far greater speeds.
1) Speed. During 15 years, I used, almost daily, an Epson QX-10
running at 4-MHz, and never felt the need for more speed. It is only
when I assembled large ASM files, that I had a problem. I remember
that my record was waiting during 15 minutes for the assembler.
2) Memory. How many programs have you written that do not fit in
memory? (If I remember well, my record is a 85KB program, using 5 or 6
overlays.)
3) Application (interprocess) protection. What? you have written a
program so badly behaving that it is running "errant"? Why don't you
use something better? Ever heard about Forth? They have been running
multi-user, multi-tasking for ages, and Forth has almost no
protection...
4) Show stopper. Indeed. You just forgot that the memory space is a
function of the "memory models" of the OS, and CP/M and CP/M-86 are,
simply, 8-bits and 16-bits Operating Systems. If you want a 32-bits
OS, what are you waiting?
> DEC VAX 1978 for the 11/780 was also 32bit. It's big thing was then an
> still the ability to protect once users program from the effects of
> another while dealing with larger memory arrays. In a multitasking,
> multiuser envronment this is a significant requirement.
You're talking, again, of memory protection (personally, I was
thinking that the VAX was better known for its virtual memory).
Anyway, this is a problem of bad programming your program. If it was
well done, you would not need memory protection.
> Not better, but it did fit the 64k and floppy envronment well. Of
> course DEC had RT11 wich fit the 16bit PDP11 with 32kW of ram
> and floppies very well. The point being small computers need small
> OSs and CP/M is a small file system and barely an OS by most
> standards.
CP/M is small: Ok. "Barely an OS": re-read the definition of "OS"...
And why would people had used this term to describe it? Were they all
idiots "30 years ago"? Finally, if you think that CP/M is too small,
how would you describe Unix Version 7...? It runs with only 40 system
calls, like CP/M. Do you need 65535 system calls to be an "OS"?
> You can communicate with mose code but sending pictures with it is
> not easily accomplised. It's the lowest common denominator for radio.
In case you do not know, it is a FUNDAMENTAL theorem of Computer
Science that any "universal" computer is able to emulate any other
universal computer. The only differences are: 1) speed, and 2) ease of
programing. CP/M provides the 6 uses of a computer, so is universal.
Everything you can do under CP/M can be done under another computer,
like an IBM Mainframe. Now, we are back to the difference between a
truck and a bicycle. When I move, I need a truck but, the rest of the
year, a bicycle is enough (unless I am a pig-headed American like you,
who considers that a Hummer is the smallest possible car: in Europe,
we have had, for 10 years, I think, something called "Smart", but I
much prefer a moped).
> I'm a fan of CP/M and still use it too. However, everything about it
> is the lowest common denominator.
Personally, I think that there is a contradiction in your sentence.
You only mention (as usual) negative terms. So, if CP/M is so bad, why
are you "a fan of CP/M" and why do you "still use it, too"? Would you
be able to proferate positive things, one time in your life?
> You will find while it was a standard it was also relatively unknown.
I know, I know... Personally, I think that the reason is that most of
the interesting stuff made by Digital Research (Personal CP/M, GSX, CP/
NET, Concurrent CP/M, Access Manager, Display Manager, Dr. Logo, DR
Graph, DR Draw) were done in 1983, that is to say: when everybody
switched like lemmings to the IBM Clown. It is typical that most of
the American "Old Timers" of this comp.os.cpm Newsgroup were users of
CP/M 2.2, not CP/M Plus, and do not know that Digital Research was
selling MAC and SID to replace ASM and DDT...
> Still bad faith.
Still stubborn. I am glad that the Atlantic Ocean separates us. Else,
would you imagine us meeting? You, only hardware; me, only software.
You, only the past; me, only the present. But you are so old that
there is no hope to change you. Anyway, now that the US are bankrupt,
you will be obliged to change, like it or not. For years, people told
me: "Why don't you go and live in the USA?" In short, 2 answers: 1)
Violence (the first time we went to the USA, we wanted to eat one of
those famous "hamburgers". I spotted one seller, went to him, ordered
some. He turned: he had a revolver at the side, just to sell
hamburgers! In England, "bobbies" are not armed. Once, in the London
"Tube", a man entered the train with a high-power hunting rifle, with
a scope... The Brits did not blink. Except those 2 times in foreign
countries, I have never seen somebody with a gun, in France. 2) Social
Security. According to the "American Dream" everybody can make it. The
problem is that there are more poor people in the USA than in Europe,
and they are much more poorer than here. Since it is highly unlikely,
due to its bankruptcy, that the USA will remedy those 2 points, I
shall remain in France. At least, the food is better.
Glen Herrmannsfeldt
(snip)
> As I wrote several years
> ago, "George Bush will end up in the garbage bin of history". I tried
> to find how many years ago I wrote this, but could not find it (it was
> in answer to Chuck Falconer).)
There is a story today that 370 artists around the country were
hired to each design and create one ornament for the white house
christmas tree. The only requirement that it had to have a
red, white, and blue theme.
A Seattle artist designed and sent in a nice red white and blue
ornament that, in very small print, said "Impeach Bush" all over it.
It seems it was the only one rejected by the white house.
The artist is well known, among artists, for mixing art and politics.
-- glen
Ole Christensen
So why put more pressure on that little Z80?
Think Netatalk *Ref.1. or Samba *Ref.2.
That are servers for "foreinger" file-system
and the processing power are done where that
power exits.
BRIDOS or the later JetNet II, (CP/M+ 3.0 comp.),
(the name could be different at Your location),
are using RS-422 and HDLC,
(there are Z8530 card for PC + driver for Linux).
In my case i run RS-422 with 4 wire + shield,
but i think it could be run in only 1 pair + shield?,
the server is a SuperJet meaning an:
Jet-80 SCB, (4 MHz 128Kb),
+ Adaptec ABC4000 SASI->ST506, (2x Wren CDC 9415-86)
+ Sysgen SC4000 SASI->QIC-02
the clients are Jet-80NC, diskless of cause,
(a "piggiboard" built into an Ampex 210 terminal),
and Jet-80FD, with 2x 1.2Mb. Teac.
So,... are RS-422 + HDLC faster than floppy?: yes,
(WS is up and running within a blink of a eye, (or 2)).
In JetNet II one "build" the client BIOS the same way
as a CP/M+ 3.0 BIOS, (source avail. (to "build" clients))...
This is, for my 2 cent, the way to go...
---
* Ref.1:
Netatalk is an implementation of the AppleTalk Protocol Suite for
BSD-derived systems. The current release contains support for
EtherTalk Phase I and II, DDP, RTMP, NBP, ZIP, AEP, ATP, PAP, ASP, and
AFP.
* Ref.2:
The Samba software suite is a collection of programs that
implements the SMB/CIFS protocol for unix systems, allowing you to serve
files and printers to Windows, NT, OS/2 and DOS clients. This protocol
is sometimes also referred to as the LanManager or NetBIOS protocol.
<ole>
Ole Christensen
1. GENERAL DESCRIPTION
----------------------
1.1. Introduction
-----------------
BRIDOS is a high-speed operating system for microcomputers. It is
distributed on a world-wide basis by PROSOFT in Sweden.
BRIDOS includes the BRISAM file-management system, which is based
on tree-structured files (B+ tree). It has been optimized for
data-processing applications in multi-user system and offers
big-computer capabilities in microcomputer format.
The major benefits of the system include:
* Faster data-processing.
* Safer data-processing.
* Logical and dynamic allocation of disk space
* More efficient use of disk space.
* Protection against obsolescence.
* Facilities for volumes or files of up to 64 M bytes.
BRIDOS and BRISAM offer unique data safety. The risk of lost data
and crashed files is virtually eliminated.
BRIDOS and BRISAM generate substantial reductions in run time or
all types of programs, from word-processing, book-keeping and
inventory management to production control and engineering.
Emulator technology
BRIDOS is designed to accept modular emulators of other operating
systems. An emulator create an environment that is for practical
purposes identical to the normal environment of a specific opera-
ting system.
The BRIDOS CP/M emulator thus loads into memory and creates a
CP/M environment. A call to CP/M is instantaneously translated
into a call to BRIDOS.
Commands in the application program are executed by the BRISAM
file-management system instead of the CP/M file manager.
A program written for CP/M runs under BRIDOS at higher speed and
with minimal risk of data loss, while the user retains access to
the wide variety of software written for CP/M.
1.2. System overview
--------------------
The operating system is the fundamental program for a microcomputer.
Without an operating system, the micro computer cannot do
its job. Microcomputer performance depends on operating system
performance. The faster the operating system, the faster the
computer works.
BRIDOS use a file-management system called BRISAM, which is
designed for faster data processing on microcomputers.
BRISAM is also designed for safer data processing. fi1es are
updated while a program is running. Updating is so fast that a
power failure or an operator error does not result in loss of
data.
Disk space is allocated dynamically in logical volumes, which
ensures more efficient utilization of disk space. The area occupied
by a file grows or shrinks as data is inserted or removed.
BRIDOS volumes are organized in a hierarchical structure which
resembles the one in the UNIX system.
BRIDOS enables microcomputer programs to be written faster, more
easily - and more economically.
The user does not have to worry about obsolescence, An 8-bit
computer and a BRIDOS operating system offer the same data-
processing capability as a 16-bit computer.
In networks and multi-user applications, data files can be shared
quickly, simply and dynamically. In the 1.52 version of BRIDOS,
as many as 16 users can access the files on a hard disk without
loss of speed are efficiency. They can also overlap on the same
file or files.
BRIDOS inc1udes a complete access control system for protection
of files. each user has a private password and can easily designate
the user or group of users who have access to his files.
Access can be be either Read only or Read/Write.
The printer spooler and de-spooler are designed for administration
of printing queues based on user priorities and codes for
paper formats. As many as 9 printers and 256 different format
queues can be supported.
BRIDOS also includes an internal electronic mail system which
can be used for high-speed transfer of messages between users.
BRIDOS works with larger files and larger volumes than any other
operating system for microcomputers. A volume or a file can be as
large as 64 M bytes.
BRIDOS uses memory space more efficiently. The user has access to
the same application program area as CP/M, with the additional
benefit of index file support.
BRIDOS includes a number of utility programs. one of these can be
used to convert CP/M file formats to BRISAM formats for faster
loading and more efficient utilization of disk space, Files can
also be accessed directly through BRISAM, for unmatched interactive
file performance.
1.2.1, Access contro1 and allocation o disk access
--------------------------------------------------
Conventional multi-user operating system for microcomputers
divide a hard disk into physical volumes. A user logs in on a
letter which represents a physical portion of the disk. The
amount of physical space reserved for this letter is constant and
does not necessarily correspond to the actual amount of data .
contained.
In other words, the space remains physically allocated, whether
it hold 4 K bytes of data or 4 M bytes.
Another problem involves the use of programs which generate
temporary files, such as WORDSTAR and COBOL compilers. The system
must include a facility for ensuring that only one user at a time
is processing a file in a given volume with such a program, and
the procedures available of doing this are general1y ineffective.
In contrast to conventional systems, BRIDOS allocates disk space
in terms of logical volumes, not physical ones. A letter does not
represent a physical portion of the disk. Instead of attaching a
letter to a user BRIDOS attaches letter to volumes, within
witch 7 levels of sub-volumes can be created.
Any user can attach any letter to any volume to which he has
access rights, The creator of a volume uses simple, straight-
forward routines to grant or deny access to other users, i.e.
Read only, Read/Write or No Access.
BRIDOS also al1ocates disk space dynamically. This means that the
space occupied by a volume increases or decreases as data is entered
or removed so that the space always corresponds to the
actual amount of data in the volume.
For example, after entering the system with his private password,
User X attaches the letter ''B:'' to DOCTXT and then logs in on
B:''. Within DOCTXT he has created 3 sub-volumes, DOCMAN, DOCLET
and DOCPRO, These volumes contain the texts of manuals letters
and current projects, respectively.
User X has assigned access rights as follows:
-- all users have access to DOCMAN
-- users Y and have Read-only access to DOCLET access to this
volume is denied to all other users except X, of course
-- user Z has Read/Write access to the files in the sub-volume
DOCPRO (and of course user X.
As the files in DOCPRO may be processed with programs that generate
temporary files, User X has also assigned exclusive access
to DOCPRO. This means that only one user at a time has access to
this volume.
As soon as a user has finished working in the DOCPRO volume, any other
authorized user has access to it automatically. Exclusive access
can easily be assigned on either a temporary or a permanent basis.
When users other than User X enter the system with their pass-
words, they can also attach ''B:'' to DOCTXT by simply writing
ATTACH B:DOCTXT.
Access to the files in any volume is assigned by its creator, so
that user logged into ''B:'' have access to the files in the
volume attached to ''B:'' in accordance with the access rights
assigned by User X.
Allocating disk space logically and dynamically provides a number
of major benefit.
First, utilization o disk space is highly optimized, as the
physical space actually occupied by a volume is a function of the
quantity a data it contain not of an arbitrary - and more or
less permanent - division of the disk.
Second, it takes much less time to locate a file in a volume.
Third, BRIDOS is more convenient than other multi-user systems.
Any user can attach a letter to any volume, independent of the
physical location of the volume on the disk, Access is defined by
the creator of the volume.
Fourth, the entire access control system is user-oriented, highly
flexible and easy to operate. The creator of a volume can assign
access in terms of a specific user, a group of users or all users.
Fifth, BRIDOS provides a simple solution to the problem of processing
a file with a program that writes temporary files.
1.2,2, Network structure
------------------------
A typical BRIDOS network is based on a central unit in which
BRIDOS resides. The hardware in this unit consists of a CPU with
a 64 KB RAM and a 40 MB Winchester drive. All data in the system
are stored here, The unit also includes up to nine processors for
application programs run by users from terminals.
File management and the data base are completely separated from
data processing. A1l file management within the system is handled
by BRISAM, the file-management system incorporated in BRIDOS.
The quantity of information in the communications network is
minimized, because the entire file-management system resides in
the central unit. BRISAM is used for all file-management.
In operation, the application program asks for information by
means of a search key. All information is retrieved by BRISAM and
then transmitted to the application-program processor,
In conventiona1 systems, the application processor must be used
for searches. It also has to handle information about the physical
location of data on the disk. The processor searches and
updates the index tree of the file-management system.
This approach has two major disadvantages:
1. Large quantities of information must be transmitted between
the master processor and the application processors.
2. The conventional system is highly sensitive to a fault in an
application processor For example, if an index file is not
updated, it will be difficult to retrieve the input data.
Each workstation in a BRIDOS network operates independently of
the others.
If one user wants a file in order to write a report, and another
user wants to process the same file statistically, BRIDOS can
serve them both without any noticeable delay.
Degradation c1ose to zero
Degradation is the concept witch describes the percentage deterioration
in the operating efficiency of a system as each user is connected
to the network.
In conventional networks and multi-user systems, a single processor is
normally used for all file management as well at all data processing.
The problem is obvious: operating speed decrease with the number
of users,
The BRIDOS multi-user system avoids this sort or crippling effect.
Each workstation is served by its own processor. A user does not
generate a load on the rest of the network until information is
requested from the central register,
Very small quantities of information are transmitted in the
BRIDOS network for search and updating of files, so that there is
no measurable degradation for data-processing applications, no
matter how many users are being served.
1.2.3. Maximum data security
----------------------------
Data security is just as important as speed and efficiency.
BRIDOS is designed for fast updating.
The system ensures that each transaction is completed in its
entirety.
The risk of updating a data file without updating the index file
is absolutely minimal.
This is the best possible protection in the event of potential
disasters such as power failure or unintentional reset,
Application software and file-management are physically separated.
A fault in an application program can never crash information registers
or interfere with the performance of other programs.
Simultaneous updating of index files and data files completely
eliminates the risk of crashed files.
1.2.4. Technical characteristic of the BRISAM fi1e-management system
--------------------------------------------------------------------
BRISAM offers high-speed index sequential search in ascending or
descending key order. A file can be as large as 64 Mb. The number
of records in a file is limited only by overall disk capacity.
A special buffering mode is used for transactions consisting of
more than one operation. Transaction time is dramatical1y reduced.
State-of-the-art technology including a B+ tree and advanced
buffer management generates extremely fast file-update operations.
BRISAM has been tested against a typical file-management system
for microcomputer. BRISAM executed INSERT, DELETE and UPDATE 5
times as fast, on search operations, BRISAM was 2-3 times as fast.
Files are fully updated at the end of each transaction, other
operating systems cannot execute updating unless the file has
been closed. If BRISAM is running and a power failure occurs, the
updated file is not lost,
A high1y optimized cache-buffering system is used internally to
minimize the number a disk accesses and guarantee data safety.
A file operation that writes to the disk is first completed
through the internal buffers, If the operation has to be aborted,
nothing is written on the disk.
Disk space is allocated dynamically. A file can be as large as 64
Mb or as small as 512 bytes.
Disk space is automatically al1ocated or deallocated in pages of
512 or 1024 bytes as a file grows or shrinks. Pages are positioned
as close as possible to the optimal position on the disk.
BRISAM ensures high efficient utilization of disk space. An
index sequential file built in random key order is packed to 85%
of capacity.
With ordered inserts - ascending or descending - the file is
packed to 100% capacity,
Files do not have to be reorganized for packing.
When records are de1eted from a file, disk space is reclaimed.
Pages are always at least half-full.
Sequential files are also supported for data with fixed or
variable record length.
A volume directory is maintained in sorted order, The directory
is accessible by index sequential search operations, using the
file name as a key,
For each volume, the name, date of creation and date of last
backup are maintained This information is presented on the
screen as soon as a user starts work on a volume.
Volumes are organized as hierarchies. A volume may contain 7
levels of sub-volumes.
BRISAM includes functions for multi-key access as well as
generation of unique keys.
BRIDOS User's Guide for PHILIPS P3000 1-6
<eof>
Steven Hirsch
> BRIDOS or the later JetNet II, (CP/M+ 3.0 comp.),
> (the name could be different at Your location),
> are using RS-422 and HDLC,
> (there are Z8530 card for PC + driver for Linux).
>
> In my case i run RS-422 with 4 wire + shield,
> but i think it could be run in only 1 pair + shield?,
> the server is a SuperJet meaning an:
> Jet-80 SCB, (4 MHz 128Kb),
> + Adaptec ABC4000 SASI->ST506, (2x Wren CDC 9415-86)
> + Sysgen SC4000 SASI->QIC-02
> the clients are Jet-80NC, diskless of cause,
> (a "piggiboard" built into an Ampex 210 terminal),
> and Jet-80FD, with 2x 1.2Mb. Teac.
>
> So,... are RS-422 + HDLC faster than floppy?: yes,
> (WS is up and running within a blink of a eye, (or 2)).
>
> In JetNet II one "build" the client BIOS the same way
> as a CP/M+ 3.0 BIOS, (source avail. (to "build" clients))...
Ole,
Can you provide a link for JetNet and BRIDOS? I don't turn up anything on
Google.
no....@no.uce.bellatlantic.net
<roch...@laposte.net> wrote:
>Allison "30 years ago..." Parent wrote:
>
>> As to the rest of the capability editing full motion video is really
>> not a posibility with CP/M or for that fact even editing a still
>> frame.
>
>But why would you want to "edit full motion video"?
Well for one reason creating animation without killing trees by the
trainload.
>And is "Bill" your first name? I answered Bill, not you.
If anything you opened the door.
> How do you call that, in American, answering at the place of someone else?
Another attempt at group insult.
>Is it something that good girls do?
Now an ad hominin attack?
>
>> It still cannot browse the internet and your still longing for a IP
>> interface to even connect to the net.
>
>It all depends what you call "browse the Internet" and, anyway, you
>have done nothing to bring TCP/IP under CP/M.
Not yet. and specially not for a PC thats been done before.
I have to solve two problems in CP/M first. One is getting it to do
two things concurrently and a suitable interface for Ethernet.
The IP stack is already in hand but running PPP over serial is
not really that useful as I could run a simpler modem program
and save effort.
>> Not used or useful under CP/M as you can expect about a megabyte
>> per page at 300 DPI and most CP/M system without a hard disk did not
>> have disks large enough to hold that much. Those scanning rigs were
>> sold for and used with DOS on PCs.. I still have one that fits in
>> place of a Epson MX80 ribbon cartridge. Worked ok for black and white
>> at low resolution and about 2-3minutes per page. Also the very
>> definition of slow and barely useful.
>
>You just forgot that scanners of the CP/M area did not made 300 DPI.
>Anyway, I have never thought that they were useful.
Mine does both 150 and 300dpi.
>> Yes, you really do understand less. speed for one, memory,
>> application( interprocess) protection. The latter is since so that a
>> errant program doesnt crash the whole show (CP/M and DOS share the
>> inability to employ that). And the show stopper is the ability to
>> have larger than 640K(1mb cor CP/M86) memory spaces to take
>> advantage of larger memeory arrays and do calulations to higher
>> precision at far greater speeds.
>
>1) Speed. During 15 years, I used, almost daily, an Epson QX-10
>running at 4-MHz, and never felt the need for more speed. It is only
>when I assembled large ASM files, that I had a problem. I remember
>that my record was waiting during 15 minutes for the assembler.
Well of you list I can pick 6 things any two and make my point.
With CP/M you cannot run the modem doing a down load and also
open and view multiplan. In short it cannot walk talk and chew gum at
the same time.
>2) Memory. How many programs have you written that do not fit in
>memory? (If I remember well, my record is a 85KB program, using 5 or 6
>overlays.)
Over a dozen applications most test programs and suites. The best was
actually small but it had to run on a system with only 8K of ram but
the application was over 40K.
>3) Application (interprocess) protection. What? you have written a
>program so badly behaving that it is running "errant"? Why don't you
>use something better? Ever heard about Forth? They have been running
>multi-user, multi-tasking for ages, and Forth has almost no
>protection...
Maybe I didn't write the program, maybe it's someone elses and I
source is not available. You can't know. However you talking about
cooperative multitasking and that may or may not be the case. For
multiuser systems it's always arisk especially is the users are
developers.
>4) Show stopper. Indeed. You just forgot that the memory space is a
>function of the "memory models" of the OS, and CP/M and CP/M-86 are,
>simply, 8-bits and 16-bits Operating Systems. If you want a 32-bits
>OS, what are you waiting?
Well it proves you know little about OSs and what was possible.
For example RT11 existed as three primary versions all ran on 16bit
hardware and in a 32kword (16x32K) one was single task like CP/M,
RT11FB could do foreground and background operations and tehre was a
batch mode, version. That same PDP11 could run RSTS a multiuser OS or
RSX11 that could both multiuser and multitask. If that wasn't enough
Unix as we know it was developed on it. So accepting a 16bit OS that
can't even print in the background (even dos could) is rather poor.
I use PDP11 as an example but I can pick from a larger list.
>> DEC VAX 1978 for the 11/780 was also 32bit. It's big thing was then an
>> still the ability to protect once users program from the effects of
>> another while dealing with larger memory arrays. In a multitasking,
>> multiuser envronment this is a significant requirement.
>
>You're talking, again, of memory protection (personally, I was
>thinking that the VAX was better known for its virtual memory).
Yes and that was implemented with a protection scheme. so that
multiple programs would not interact in bad ways. Specifically any
program is safely contained in its own space so the OS can continue if
there is an error.
>Anyway, this is a problem of bad programming your program. If it was
>well done, you would not need memory protection.
Thats both foolish and short sighted. On a machine with more than one
application or user a bug in any program can kill the whole system.
If you can trust all the programs running concurrently great but the
more there are the mode tlikely an interaction and kill the whole
show. The classic CP/M crash is forgetting to replace the system disk
and the OS dies.
>> Not better, but it did fit the 64k and floppy envronment well. Of
>> course DEC had RT11 wich fit the 16bit PDP11 with 32kW of ram
>> and floppies very well. The point being small computers need small
>> OSs and CP/M is a small file system and barely an OS by most
>> standards.
>
>CP/M is small: Ok. "Barely an OS": re-read the definition of "OS"...
I did and barely an OS does not say it is not an OS! I am saying it
has all of the qualifications at the most minimal level.
>And why would people had used this term to describe it? Were they all
>idiots "30 years ago"? Finally, if you think that CP/M is too small,
>how would you describe Unix Version 7...? It runs with only 40 system
>calls, like CP/M. Do you need 65535 system calls to be an "OS"?
It's not how many systems calls but the need to program around those
that are ineffective or broken. Example is a lot of the modem
programs for CP/M as there is only one device that has status checking
for RX and TX and thats the console. That means the console must be
shared with a modem (read possible dropped characters) or direct IO
so the modem has it's IO and can still comminicate with console.
Makes running modem at 9600 or faster a real problem, some systems
like the Kaypro have a rough time at 2400.
>
>> You can communicate with mose code but sending pictures with it is
>> not easily accomplised. It's the lowest common denominator for radio.
>
>In case you do not know, it is a FUNDAMENTAL theorem of Computer
>Science that any "universal" computer is able to emulate any other
>universal computer. The only differences are: 1) speed, and 2) ease of
>programing. CP/M provides the 6 uses of a computer, so is universal.
Correct. But your challenge was and still is name things that linux
or any other OS can do that CP/M can't and the answer is two things at
once or make the appearance of that. The fact that even a Turing
machine with the right OS can suggests you miss the point.
CP/M is not universal. It was ported to a limited set of processors
(8080/8085/z80, 68000, Z8000). If I use your universal I'd then have
to say Linix is as it's been ported to more differernt CPU
archetectures (ARM, Intel x86, Alpha, 68000, dragon ball, Coldfire,
MIPS, RISC PC, SUN SPARC, atmel AVR32, IBM power archetecture,
IBM RS6000 series, IBM390[31bit], StrongARM and many more).
>Everything you can do under CP/M can be done under another computer,
>like an IBM Mainframe. Now, we are back to the difference between a
>truck and a bicycle. When I move, I need a truck but, the rest of the
>year, a bicycle is enough
When you want to go fast or carry more than a water bottle a truck,
car or even a trailer on the bicycle is maybe the better way. Of
course you insist the fastest way from Rome to Paris is a bicycle
you are being foolish.
>(unless I am a pig-headed American like you,
>who considers that a Hummer is the smallest possible car: in Europe,
>we have had, for 10 years, I think, something called "Smart", but I
>much prefer a moped).
That is both an insult and rude and that is apparently your style.
But if you must know I dispise Hummers and feel that people that
purchase them are both stupid and wasteful. Even your mercedes
is far more wasteful than my set of wheels.
There is one difference, you call me names insult me and others if you
don't like our opinions. If we annoy you your response is to start
calling us "pigheaded Americans" and other attacks and going off on
tangents.
>> I'm a fan of CP/M and still use it too. However, everything about it
>> is the lowest common denominator.
>
>Personally, I think that there is a contradiction in your sentence.
>You only mention (as usual) negative terms. So, if CP/M is so bad, why
>are you "a fan of CP/M" and why do you "still use it, too"? Would you
>be able to proferate positive things, one time in your life?
Becuase over the years you haven't listened once. I use it beacuse it
does work, it is the lowest denominator and yes thats not a
contradiction. My use and enjoyment of it is because of it's
simplicity and that it is in fact the lowest common denominator. You
call it negative, I call it simply what it is and there is no
negativity in that. Being simple is good, and it's is certainly
useful. But to make foolish claims that CP/M is a world beater is
indeed silly. There are tasks it does not do and many are fairly
simple. Some are fixable and some or far out of it's scope of
operation.
But at the end of the day I'll use CP/M for one thing and in a
heartbeat use Apple OSX or Linux to do something that is far out of
the reach of CP/M. I will say CP/M is not almighty and just as fast
say I can find manything in LInux, VMS, OSX and many others I work
with and use that I find dumb and annoying. I will not ever deny CP/M
is not the last word and linux is not the last word either.
>> You will find while it was a standard it was also relatively unknown.
>
>I know, I know... Personally, I think that the reason is that most of
>the interesting stuff made by Digital Research (Personal CP/M, GSX, CP/
>NET, Concurrent CP/M, Access Manager, Display Manager, Dr. Logo, DR
>Graph, DR Draw) were done in 1983, that is to say: when everybody
>switched like lemmings to the IBM Clown. It is typical that most of
>the American "Old Timers" of this comp.os.cpm Newsgroup were users of
>CP/M 2.2, not CP/M Plus, and do not know that Digital Research was
>selling MAC and SID to replace ASM and DDT...
>
>> Still bad faith.
>
>Still stubborn. I am glad that the Atlantic Ocean separates us. Else,
>would you imagine us meeting? You, only hardware; me, only software.
>You, only the past; me, only the present. But you are so old that
If you met me you'd find I'm about the future. What I do elsewhere is
not reported much here. The fact that I'm part of a team developing
on A Blackfin DSP is not backward or even old. I do enjoy software
and hardware to limit myself to either is half a project.
>there is no hope to change you. Anyway, now that the US are bankrupt,
Another ad hominin attack. My age is not a limit to me or those far
younger. Expereince is value and I bring it and hope to learn more.
From you I learn stubborn belief, insults, attacks and rudeness but
nothing technically new or even previously valuable.
>you will be obliged to change, like it or not. For years, people told
>me: "Why don't you go and live in the USA?" In short, 2 answers: 1)
>Violence (the first time we went to the USA, we wanted to eat one of
>those famous "hamburgers". I spotted one seller, went to him, ordered
>some. He turned: he had a revolver at the side, just to sell
>hamburgers! In England, "bobbies" are not armed. Once, in the London
>"Tube", a man entered the train with a high-power hunting rifle, with
>a scope... The Brits did not blink. Except those 2 times in foreign
>countries, I have never seen somebody with a gun, in France. 2) Social
>Security. According to the "American Dream" everybody can make it. The
>problem is that there are more poor people in the USA than in Europe,
>and they are much more poorer than here. Since it is highly unlikely,
>due to its bankruptcy, that the USA will remedy those 2 points, I
>shall remain in France. At least, the food is better.
Sigh, again with the insults. So little you know and so little you
wish to hear.
I guess reminding you that the British, Candians and the free French
allowed you to continue speaking French and not German is pointless.
>Yours Sincerely,
>Mr. Emmanuel Roche, France
Sincere, I doubt.
Allison
Jeff Jonas
>> > peripheral chips are used, and networking makes no exception here.
>> > A Z80 running CP/M would profit a lot from
>> > offloading the network stack to a network controller
>> > and such things exist already.
Computer architecture always fascinated me, particularly I/O systems.
My joke (mostly to myself)
is how the mainframe folks will "have their day".
Mainframes never had the fastest CPUs but achieved high thruput
with a heirarchy of parallel I/O processors and specialized controllers.
Minicomputers kinda approximated that with smart I/O controllers
(when available; most were not-so-smart).
On a micro level, the Z80 CPU offloaded some tasks
to its peripherals, such as CTC auto-reloading,
parallel I/O chips being programmed
to interrupt on only certain conditions.
The strength of the "classic" Z80 CPU
was the variety of peripherals
- quad CTC: Counter Timer
- dual PIO: parallel I/O
- dual SIO: sync serial I/O ports
- dual UART: async serial I/O ports
They all interfaced with just address decoding,
self-arbitrating interrupts using the vectored mode 3.
The SIO chip was kinda an early NIC when used in SDLC mode:
up to 1 MegaBit speed, understoond the frame format
for responding to its own address,
automatic CRC generation and checking, etc.
I am led to believe that AppleTalk was based on the Z80 SIO/8530.
Many protocol converts were just a Z80 CPU
and a bank of SIO chips for the sync protocol.
>> At some point, the Z80 ceases to be your CPU and becomes instead
>> merely a PPU - peripheral processing unit.
In systems of the 80s, there tended to be Z80s in all roles.
Zilog's own MCZ2/50 used a Z80 for the main CPU
and another in the floppy controller to offload command processing.
The Zilog System 8000 used the Z8000 CPU
but several Z80s as firmware-programmed dedicated controllers,
or just the Z80 chips kinda shoe-horned to the Z8000
(I remember the device driver having to force
the Z80's "re" "ti" onto the peripheral bus
to reset the Z80 peripherals interrupt-pending).
Even the terminals and printers of the time
were microcontrollers such as the Z8.
Michael:
>Well, it does what its best at. A Z80 can't interface a HD floppy
>directly very well, so it uses a FDC and a DMA controller.
That's a nice distinction of microprocessors (Z80)
vs. microcontrollers (Z8, PIC, AVR, ...).
I really "missed the boat" with the Z8 and PIC
by looking at the spec sheet and thinking
"that's really WEAK compared to the Z80"
without understanding its role in the entire system.
Microcontrollers are happiest running small programs
from their internal ROM (flash, EPROM, ROM)
with re-programming only for updates
whereas microprocessors are happiest with lotsa external RAM
and dynamically loading various tasks (from mass storage:
floppy, hard drive, flash drive, network ...).
>Steal cycles from its RAM for video, and performance is bad,
I hope you mean "real cycle stealing" as in using
the refresh cycles for read/write of static RAM,
accessing data RAM during opcode fetch from a different RAM bank, etc.
All without impacting the CPU's cycles at all.
I ask because the IBM system 1130's manuals referred to
cycle stealing but it halted the CPU, so it was really DMA.
> so you use a video controller with private RAM.
Or dual ported ram.
> The SIO has a small FIFO buffer to reduce
>latency requirements, and so on. In the eighties, all of that was already
>common technology, and had people had ethernet back then, they had built
>network controllers.
As I noted, the SIO was capable of NIC-style operation in sync mode
with filtering frames by address and only accepting those
with the station's address, or broadcasts.
If it got lost or confused, it could "hunt" for the next frame
without any CPU intervention.
> The Z80 family includes many peripheral chips for a reason.
> These days, PICs and the like allow to build you own
> peripheral controllers at low cost and using low power.
Kinda, maybe, almost.
The PIC UARTS don't have the deep FIFO of the SIO or DART,
depending on interrupt servicing to make up for the lack of hardware.
The more I read about the PIC 18's peripherals such as USB,
the more I'm frustrated with the lack of silicon support.
Yes, they're cheap, single chip solutions
but it bothers me that we're using CPU cycles
for the lack of silicon for FIFOs, protocol state machines, etc.
But proponents would say
"why not, that makes them more versatile and CPU cycles are cheap".
I guess I can't totally ignore my "vintage" feelings:
(stolen from The Life Of Brian)
every byte is precious
every byte is great
when a byte is wasted
DEC gets quite irate
Mr Emmanuel Roche, France
Many, many thanks for sharing with us a little knowledge!
It is the most interesting thin I have read on the comp.os.cpm
Newsgroup for years.
(It reminds me of PICK.)
The only thing I could find is:
http://www.old-computers.com/museum/computer.asp?st=1&c=904
So, it seems that you have a real rarity...
Maybe you could setup a Web page dealing with it?
Mr Emmanuel Roche, France
Personally, since I am normal, I am starting to tire of talking to a
wall.
Reading your ramblings, I noticed something:
> With CP/M you cannot run the modem doing a down load and also
> open and view multiplan. In short it cannot walk talk and chew gum at
> the same time.
> Well it proves you know little about OSs and what was possible.
> For example RT11 existed as three primary versions all ran on 16bit
> hardware and in a 32kword (16x32K) one was single task like CP/M,
> RT11FB could do foreground and background operations and tehre was a
> batch mode, version. That same PDP11 could run RSTS a multiuser OS or
> RSX11 that could both multiuser and multitask. If that wasn't enough
> Unix as we know it was developed on it. So accepting a 16bit OS that
> can't even print in the background (even dos could) is rather poor.
> I use PDP11 as an example but I can pick from a larger list.
> Yes and that was implemented with a protection scheme. so that
> multiple programs would not interact in bad ways. Specifically any
> program is safely contained in its own space so the OS can continue if
> there is an error.
> Thats both foolish and short sighted. On a machine with more than one
> application or user a bug in any program can kill the whole system.
> If you can trust all the programs running concurrently great but the
> more there are the mode tlikely an interaction and kill the whole
> show. The classic CP/M crash is forgetting to replace the system disk
> and the OS dies.
> It's not how many systems calls but the need to program around those
> that are ineffective or broken. Example is a lot of the modem
> programs for CP/M as there is only one device that has status checking
> for RX and TX and thats the console. That means the console must be
> shared with a modem (read possible dropped characters) or direct IO
> so the modem has it's IO and can still comminicate with console.
> Makes running modem at 9600 or faster a real problem, some systems
> like the Kaypro have a rough time at 2400.
That is to say: all your remarks concern CP/M 2.2... You simply do not
seem to know that other versions of CP/M existed (CP/M Plus, Personal
CP/M, CP/M-86 Plus, Concurrent CP/M, Concurrent DOS). As I explained
several times, Digital Research intended CP/M Plus to be multi-
tasking. But, when the Z-800 failed to appear in time, they launched
the 8-bits version without multi-tasking. (Remember: CP/M Plus is a
single user version of MP/M-II.) Now, the 16-bits version, CP/M-86
Plus, is multi tasking... You can have up to 8 programs running at the
same time in the TPA. And the "segments" of those programs are in
separate locations, contrary to MS-DOS. There was even a "common code"
option, so that those 8 programs would only load their data segments
in the TPA, and use the same code segment for them all.
Also, WordStar was multi-tasking: it was printing while you were
editing. It is not too much difficult to have a program handling 2
tasks at the same time. Surely, someone like you would be able to do
that? (Hint: I published in the comp.os.cpm Newsgroup a program that
was managing 8 "things"...)
About your bombing with PDP-11 stuff: it is the same with CP/M. CP/M
was a FAMILY of Operating Systems. The only one you know, CP/M 2.2,
was single user, single tasking. But MP/M-II was multi user, multi
tasking with a Z-80 running at 4-MHz! As for Concurrent, the last
version was able to manage up to 64 terminals at the same time, with
VGA graphics. How would that have been possible, without multi
tasking?
I was forgetting: Concurrent has a 256-level priority scheme (not
protection). It was enough for it to run even MS-DOS programs.
Now, for another remark:
> CP/M is not universal. It was ported to a limited set of processors
> (8080/8085/z80, 68000, Z8000). If I use your universal I'd then have
> to say Linix is as it's been ported to more differernt CPU
> archetectures (ARM, Intel x86, Alpha, 68000, dragon ball, Coldfire,
> MIPS, RISC PC, SUN SPARC, atmel AVR32, IBM power archetecture,
> IBM RS6000 series, IBM390[31bit], StrongARM and many more).
I wonder if you are joking, or are simply simple minded? "Universal"
in Computer Science texts does not mean "ported to every processor";
it means "able to solve all problems". (Forth is much more "universal"
than Linux, in your sense, since it runs on the RCA 1802.)
> Of course (if) you insist (that) the fastest way from Rome to Paris is a bicycle
> you are being foolish.
It is you, who is foolish. In Europe (and France in particular), we
have a dense network of trains... Remember, those steam locomotives?
Well, they are a thing of the past. Now, we use electricity. In
France, we have the TGV (Train à Grande Vitesse), a train running at
400 kilometres per hour (the speed of a non-jet airplane) or, roughly,
4 times the speed limit on Motorways. From the city center to the next
city center. But, again, you cannot understand, since in the USA
people are living in the suburbs, instead of city centers. 4 times per
year, I walk to the railway station, board a train, and go and visit
relatives. No need for a bicycle, moped, or car. (At the destination,
walking is enough to go to their homes.) All that with electricity.
> There is one difference, you call me names insult me and others if you
> don't like our opinions. If we annoy you your response is to start
> calling us "pigheaded Americans" and other attacks and going off on
> tangents.
No, no, not "others"... Only YOU! You are my all-time favorite pig-
headed American... (I particularly liked the "going off on tangents",
since it is YOU who who, in the comp.os.cpm Newsgroup, keep mentioning
the PDP-11 and things that are obviously outside its realm, like "full
motion video"...)
> Sigh, again with the insults. So little you know and so little you
> wish to hear.
And you, All Knowingly, would you be able to write all this in French?
> I guess reminding you that the British, Candians and the free French
> allowed you to continue speaking French and not German is pointless.
Yes, because you are off topic. As usual...
>Yours Sincerely,
>Mr. Emmanuel Roche, France
> Sincere, I doubt.
> Allison
If I did not enjoy writing you, I would not have writen it...
no....@no.uce.bellatlantic.net
<roch...@laposte.net> wrote:
>As usual, Allison "30 years ago..." Parent wants to have the last say.
>
Give it up.
>Personally, since I am normal, I am starting to tire of talking to a
>wall.
I have for a while.
Well your still wrong. I do have CP/M in many flavors and MP/M too.
Concurrent CP/m and Concurrent dos are not CP/M 80. That like saying
DOS and Windows NT4 are the same, sure they can run the same
programs but their internal structure are very different. They were
also far from mainstream.
>Also, WordStar was multi-tasking: it was printing while you were
>editing. It is not too much difficult to have a program handling 2
>tasks at the same time. Surely, someone like you would be able to do
>that? (Hint: I published in the comp.os.cpm Newsgroup a program that
>was managing 8 "things"...)
Chest pounding likely a useless tool too.
>
>About your bombing with PDP-11 stuff: it is the same with CP/M. CP/M
>was a FAMILY of Operating Systems. The only one you know, CP/M 2.2,
So was RT. Again with the errors. I always address CP/M80 here as
the larger population are 8bit. Your a minority.
>was single user, single tasking. But MP/M-II was multi user, multi
>tasking with a Z-80 running at 4-MHz! As for Concurrent, the last
>version was able to manage up to 64 terminals at the same time, with
>VGA graphics. How would that have been possible, without multi
>tasking?
MP/M is not CP/M.
>
>I was forgetting: Concurrent has a 256-level priority scheme (not
>protection). It was enough for it to run even MS-DOS programs.
Right its not priority. Cuncurrent was created to get over the
problem os single tasking on PC that were starting to get fast enough
to be useful for multitasking.
>Now, for another remark:
>
>> CP/M is not universal. It was ported to a limited set of processors
>> (8080/8085/z80, 68000, Z8000). If I use your universal I'd then have
>> to say Linix is as it's been ported to more differernt CPU
>> archetectures (ARM, Intel x86, Alpha, 68000, dragon ball, Coldfire,
>> MIPS, RISC PC, SUN SPARC, atmel AVR32, IBM power archetecture,
>> IBM RS6000 series, IBM390[31bit], StrongARM and many more).
>
>I wonder if you are joking, or are simply simple minded? "Universal"
>in Computer Science texts does not mean "ported to every processor";
>it means "able to solve all problems". (Forth is much more "universal"
>than Linux, in your sense, since it runs on the RCA 1802.)
I used the dictionary meaning and universal means its applied
everywhere. CP/M is not. Also you have never produced or proved the
1802 version ever existed. If you going to make foolish statments
there is a port of CP/M to VAX using CP/M 68 code. It;s also mostly
useless.
>
>> Of course (if) you insist (that) the fastest way from Rome to Paris is a bicycle
>> you are being foolish.
>
>It is you, who is foolish. In Europe (and France in particular), we
>have a dense network of trains... Remember, those steam locomotives?
Yes, the trains are very good but they are not bicycles.
>Well, they are a thing of the past. Now, we use electricity. In
>France, we have the TGV (Train à Grande Vitesse), a train running at
>400 kilometres per hour (the speed of a non-jet airplane) or, roughly,
>4 times the speed limit on Motorways. From the city center to the next
>city center. But, again, you cannot understand, since in the USA
>people are living in the suburbs, instead of city centers. 4 times per
>year, I walk to the railway station, board a train, and go and visit
>relatives. No need for a bicycle, moped, or car. (At the destination,
>walking is enough to go to their homes.) All that with electricity.
Again bashing America.
>
>> There is one difference, you call me names insult me and others if you
>> don't like our opinions. If we annoy you your response is to start
>> calling us "pigheaded Americans" and other attacks and going off on
>> tangents.
>
>No, no, not "others"... Only YOU! You are my all-time favorite pig-
>headed American... (I particularly liked the "going off on tangents",
>since it is YOU who who, in the comp.os.cpm Newsgroup, keep mentioning
>the PDP-11 and things that are obviously outside its realm, like "full
>motion video"...)
Not likely. You did the tanget thing and open the door I use that as
a opportunity to bring in everything youare apprently ingnorant of.
>
>> Sigh, again with the insults. So little you know and so little you
>> wish to hear.
>
>And you, All Knowingly, would you be able to write all this in French?
No, never wanted to learn that one. Would you be able to listen in
morse?
>> I guess reminding you that the British, Candians and the free French
>> allowed you to continue speaking French and not German is pointless.
>
>Yes, because you are off topic. As usual...
Ok, Mr Roche, you started with bashing americans and a slew of other
comments, remarks and slurs, You still resort to abuse and rudeness.
You do not like it when it is done to you and you think any one else
should accept your trash. No you are off topic and far from relevent.
Bet it known in the Quaker way this man is shunned. He is being
shunned as his behavior does not give to the greater good and is
destructive. You have no voice.
Allison
Jonathan Graham Harston
CBFalconer wrote:
> Simply emulate a Hitachi 64180, which I believe is identical to a
> Zilog Z180. It has a 1 megabyte address range, and is fully Z80
> compatible.
Talking of which, I've been adding '180 and '280 functionality to
my Z80 emulator and I'd like to try banked CPM 3 on it. I spent a
couple of hours last week roaming the web looking for a) CPM 3 and
b) how to implement a suitable BIOS, but couldn't find anything.
Any pointers?
Ta.
--
J.G.Harston - j...@arcade.demon.co.uk - mdfs.net/User/JGH
05:10:36, 18-Nov-2008 - RISC OS time rolls over to &5000000000
Ole Christensen
> Hello, Ole!
>
> Many, many thanks for sharing with us a little knowledge!
>
> It is the most interesting thin I have read on the comp.os.cpm
> Newsgroup for years.
Hello Emmanuel,
Glad to entertain You for a while!
> (It reminds me of PICK.)
Please let me learn more!
> The only thing I could find is:
>
> http://www.old-computers.com/museum/computer.asp?st=1&c=904
yes,... to bad!
> So, it seems that you have a real rarity...
could be so, we will see...
> Maybe you could setup a Web page dealing with it?
Web page, me?, no way, You joke with me!
<ole>
Ole Christensen
> Ole,
>
> Can you provide a link for JetNet and BRIDOS? I don't turn up anything
> on Google.
Steven then Your Google is no better than my!
What i found i found upstairs, and that is:
* BRIDOS Operating system version 1.52
user's guide and interface manuals
for Philips P3000
march 9. 1987
(100 p.)
* BRIDOS Operating system version 1.52
user's guide and interface manuals
Prosoft january 10. 1984
(100 p.)
* Jet-Net II Operators handbook, (50 p. Swedish)
* Jet-Net II Programmers Manual (100 p.)
and i know there should be an:
* Jet-Net II Installations handbook, (short, Swedish)
somewhere...
(all for version 1.4)
+ some floppy's.
(and the hardware to run it)...
So there we are!
sorry that i can't point You to anything better
than this, but...
<ole>
Mr Emmanuel Roche, France
> > (It reminds me of PICK.)
>
> Please let me learn more!
What, you don't know PICK? Many years ago, when I was a COBOL
programmer on IBM Mainframes, there was something called PICK which
was quite popular for companies interested principally into databases.
Searching for something that you could read, I found:
http://en.wikipedia.org/wiki/Pick_operating_system
> > Maybe you could setup a Web page dealing with it?
>
> Web page, me?, no way, You joke with me!
No, no, I am dead serious: you already have the manuals and the
floppies...
"All you have to do" is scan all the pages of the manuals, and
transfer the files of all the floppies...
Courage! This way, you will not notice those long Winter nights, you
lucky guy!
(Under CP/M, a library of ISAM functions, linkable by Digital Research
compilers, is named "Access Manager". It is little known because it
was released in 1983. It was, like most DRI stuff, running under CP/M,
CP/M-86, and MS-DOS.)
Mr Emmanuel Roche, France
>
> Give it up.
Hahaha! You don't know me!
> >Also, WordStar was multi-tasking: it was printing while you were
> >editing. It is not too much difficult to have a program handling 2
> >tasks at the same time. Surely, someone like you would be able to do
> >that? (Hint: I published in the comp.os.cpm Newsgroup a program that
> >was managing 8 "things"...)
>
> Chest pounding likely a useless tool too.
Ok. So, the first argument was that you wrote that CP/M could not do
several things at the same time. I cited WordStar. (Ever heard about
it?) Since then, I have remembered that "MTBASIC" existed... Guess
what "MT" means in its name? "Multi-Tasking"... BASIC under CP/M... I
am, simply, amazed that you do not know that it is possible to do
multi-tasking inside a program, without relying upon an Operating
System. For the simple thing that you mentioned (that I would rather
call "redirection", CP/M Plus was doing it naturally) and, since CP/M
has only 5 devices, I think that a 4-MHz Z-80 would have no problem to
do it, since MP/M-II was able to do it. Of course, the speed of the I/
O subsystem will be a limiting factor.
> MP/M is not CP/M.
Hum... What to say? MP/M is upper compatible with CP/M, able to run
all its programs unmodified. Personally, I am much more impressed by
MP/M-II than by your beloved CP/M 2.2.
> >I wonder if you are joking, or are simply simple minded? "Universal"
> >in Computer Science texts does not mean "ported to every processor";
> >it means "able to solve all problems". (Forth is much more "universal"
> >than Linux, in your sense, since it runs on the RCA 1802.)
I stand by my words.
> > (...) But, again, you cannot understand, since in the USA
> >people are living in the suburbs, instead of city centers. (...)
>
> Again bashing America.
??? Are you still of bad faith? Europeans and Americans live
differently. This is a fact, not a criticism (of what?).
I confirm: You are my all-time favorite pig-headed American... Too bad
you were not President, instead of George "War!" Bush... With somebody
like you, the USA would have been officially declared "Evil Empire"...
> No, never wanted to learn that one. Would you be able to listen in
> morse?
Yes, when I was a Boy Scout!
> Ok, Mr Roche, you started with bashing americans and a slew of other
> comments, remarks and slurs, You still resort to abuse and rudeness.
> You do not like it when it is done to you and you think any one else
> should accept your trash. No you are off topic and far from relevent.
Ok, Allison "30 years ago..." Parent, YOU started by saying that CP/M
could not send/receive a file while doing something else. I provide
the names of 2 programs which used to do precisely this thing under CP/
M 2.2. I never started bashing Americans: you are my favorite pig-
headed American. (Unless you pretend to be 350 millions of Americans,
all by yourself...) I am not rude: I did not insult you (having
received a good education, my slang vocabulary is very weak, contrary
to you.) It is you who do not stand being proved wrong.
> Bet it known in the Quaker way this man is shunned. He is being
> shunned as his behavior does not give to the greater good and is
> destructive. You have no voice.
I had to look up in my English dictionaries for "shun". It is not a
verb that I use a lot. Speaking of "behavior not giving to the greater
good", could you remind me who wrote and published more than 20
programs on the comp.os.cpm Newsgroup, while someone did not
contribute anything during 10 years? Is this "destructive"? If so, I
understand that you quake...
Mr Emmanuel Roche, France
that I never received a message from you, answering the question that
I asked when sending you the WS4 file.
So, re-send me a messsage to this spam e-mail address, so I can send
you my private e-mail address.
Mr Emmanuel Roche, France
Guess what I have found on the Internet?
1 ' This demonstrates the multitasking capability of MTBASIC
2 ' by allowing you to set the reschedule interval of multiple
3 ' LED flashing tasks. While the LEDs are flashing an
4 ' incrementing counter will show on your terminal.
Who said that a program under CP/M 2.2 could not multitask?
Yours Sincerely,
Mr. Emmanuel Roche, France
10 INTEGER X,Y,Z
20 INTEGER T1,T2,T3,T4,T5,T6,T7,T8,T9
310 Y=1
320 FOR X=1 TO 8
330 PRINT " ENTER RESCHEDULE COUNT FOR THE LED THAT IS SHINING"
332 PRINT " (A VALUE FROM 1 TO 50 IS RECOMMENDED) "
335 PRINT "THIS IS TASK #";X
340 PTAOUT Y
350 INPUT Z
360 START X,Z
370 Y=Y * 2
380 NEXT X
410 REM OUTPUT TO THE TERMINAL WHILE LEDS ARE FLASHING
420 FOR X=1 TO 500
430 print x
440 wait 2
450 NEXT X
460 GOTO 420
1000 TASK 1
1010 T1= PTAIN
1020 PTAOUT BXOR(T1,1)
1040 EXIT
1100 TASK 2
1110 T2= PTAIN
1120 PTAOUT BXOR(T2,2)
1140 EXIT
1200 TASK 3
1210 T3= PTAIN
1220 PTAOUT BXOR(T3,4)
1240 EXIT
1300 TASK 4
1310 T4= PTAIN
1320 PTAOUT BXOR(T4,8)
1330 EXIT
1400 TASK 5
1410 T5= PTAIN
1420 PTAOUT BXOR(T5,16)
1440 EXIT
1500 TASK 6
1510 T6= PTAIN
1520 PTAOUT BXOR(T6,32)
1540 EXIT
1600 TASK 7
1610 T7= PTAIN
1620 PTAOUT BXOR(T7,64)
1640 EXIT
1700 TASK 8
1710 T8= PTAIN
1720 PTAOUT BXOR(T8,128)
1740 EXIT
no....@no.uce.bellatlantic.net
CP/M has nothing to do with that functionality, doesnt even have to
be there.
Ole Christensen
>
>>> (It reminds me of PICK.)
>> Please let me learn more!
>
> What, you don't know PICK? Many years ago, when I was a COBOL
> programmer on IBM Mainframes, there was something called PICK which
> was quite popular for companies interested principally into databases.
>
> Searching for something that you could read, I found:
>
> http://en.wikipedia.org/wiki/Pick_operating_system
Thank You, no that was new to me...
>>> Maybe you could setup a Web page dealing with it?
>> Web page, me?, no way, You joke with me!
>
> No, no, I am dead serious: you already have the manuals and the
> floppies...
>
> "All you have to do" is scan all the pages of the manuals, and
> transfer the files of all the floppies...
but,... i don't have anywhere to put it in cypher-space...
> Courage! This way, you will not notice those long Winter nights, you
> lucky guy!
thats right, yes,... but, for this winter i do have some other
small "projects" beside from trying to restore software from:
Wave Mate Bullet, (CP/M+ 3.0 + utility's),
Jet-80, (CP/M+ 3.0 + utility's),
RC-Partner 750,
all 3 "hard-to-find" on "the net"...
+ all kind of software where BRIDOS & Jet-Net II are just 2.
(and it will depend on interest, from me,... and others)...
and to those "other small projects", (toys), are:
ZiLOG eZ80 Acclaim, (could be of some interest to You),
WIZnet WIZ810MJ, (could very well be of "some interest" to You),
just to name a couple, so when i find out witch end
are up and down on that "spade" i will "dig"...
(and the winter are fare to short for all!)...
> (Under CP/M, a library of ISAM functions, linkable by Digital Research
> compilers, is named "Access Manager". It is little known because it
> was released in 1983. It was, like most DRI stuff, running under CP/M,
> CP/M-86, and MS-DOS.)
yes, there are a lot of nice thing in there, sure...
<ole>
joe
> Re-Hello, Allison "30 years ago..." Parent!
>
> Guess what I have found on the Internet?
>
> 1 ' This demonstrates the multitasking capability of MTBASIC
> 2 ' by allowing you to set the reschedule interval of multiple
> 3 ' LED flashing tasks. While the LEDs are flashing an
> 4 ' incrementing counter will show on your terminal.
>
> Who said that a program under CP/M 2.2 could not multitask?
>
Is this an example of multitasking, or multiple threads in a single task?
What features in CP/M actually support multitasking? How does the OS do a
context switch?
You've shown only examples of multiple threads in a single task. Much
different than an OS that multitasks.
no....@no.uce.bellatlantic.net
>Mr Emmanuel Roche, France wrote:
>
>> Re-Hello, Allison "30 years ago..." Parent!
>>
>> Guess what I have found on the Internet?
>>
>> 1 ' This demonstrates the multitasking capability of MTBASIC
>> 2 ' by allowing you to set the reschedule interval of multiple
>> 3 ' LED flashing tasks. While the LEDs are flashing an
>> 4 ' incrementing counter will show on your terminal.
>>
>> Who said that a program under CP/M 2.2 could not multitask?
>>
>
>Is this an example of multitasking, or multiple threads in a single task?
If you ask me it's more like cooperative coroutines. The acid test is
what happens if either routine fails for any reason?
Anything under CP/M could multitask but, CP/M does not support it, it
doesn't hinder it either. In most applications cases CP/M serves as a
program loader a set of APIs for a file system and basic cooked
and raw IO. The user interface the CCP is even replaceable and
often was with CCP+ or ZCPR.
>What features in CP/M actually support multitasking? How does the OS do a
>context switch?
None, nada, there is no context switch in CP/M2.2, CP/M3 or CP/M86,
no task manager, no job/process table and the BDOS is not safely
reentrant. Single threaded, single user,single task.
>You've shown only examples of multiple threads in a single task. Much
>different than an OS that multitasks.
A real example would be to use a timer interrupt to trigger a call to
a task manager then perform a task switch. CP/M does not provide any
support for that but it can be wired into the bios to allow things
like a print spooler to run as a background task. A very basic
example fo cooperative multitasking if the BIOS supports it. There
were several products that did that by looking at the BIOS calls to
console status and stealing a few cycles to process a buffer to the
printer.
CP/M is single task. DRI has other products that multitask but they
are not CP/M though I have nothing bad to say for MP/M or Concurrent
CP/M save for they aren't CP/M.
Allison
Greegor
done by various licensees including NEC.
The point was made that some didn't even
include the text Z 80 on the markings.
This shook loose an old detail stuck in my head.
Way back in the early days of building
Taiwan XT clones, there was a drop-in
substitute for 8088 CPUs that gave a slight
performance advantage. The NEC V.20 chip.
But a lesser known detail was that it had
some kind of function for Z-80 code execution.
Did ANYTHING ever use that Z-80 functionality?
Were those 8088 work alikes ever used
in the place of Z-80 chips in single board
or embedded computers?
Did anybody ever do a CP/M 2.2 CBIOS for an
XT clone with a V.20 chip?
I think there were 8MHz versions, maybe more.
Was it a nice idea that just got overlooked?
no....@no.uce.bellatlantic.net
wrote:
>I was Googling through photos of the old Z-80 chips
>done by various licensees including NEC.
>The point was made that some didn't even
>include the text Z 80 on the markings.
>
>This shook loose an old detail stuck in my head.
>
>Way back in the early days of building
>Taiwan XT clones, there was a drop-in
>substitute for 8088 CPUs that gave a slight
>performance advantage. The NEC V.20 chip.
>
>But a lesser known detail was that it had
>some kind of function for Z-80 code execution.
Because it didn't the V20 had 8080 execution mode.
>
>Did ANYTHING ever use that Z-80 functionality?
22nice I think to run older 8080 programs.
>
>Were those 8088 work alikes ever used
>in the place of Z-80 chips in single board
>or embedded computers?
Yes, as the native mode was 8088 16bit.
Since it was also available in CMOS it was used for
embedded systems but again as 8088.
>Did anybody ever do a CP/M 2.2 CBIOS for an
>XT clone with a V.20 chip?
>I think there were 8MHz versions, maybe more.
I have a 12mhz part. But no again on the bios that I know of.
The CPU ran in 8088 mode and on instruction execution it
would run as 8080. So for the PCs that had it it really was a
infrequently used mode.
>
>Was it a nice idea that just got overlooked?
No. 8080 is not z80 and the somewhat lower performance of 8080
was perceived as not very useful.
Allison
Herb Johnson
> Mr Emmanuel Roche, France skrev:
> >>> Maybe you could setup a Web page dealing with it?
> >> Web page, me?, no way, You joke with me!
>
> > No, no, I am dead serious: you already have the manuals and the
> > floppies...
>
> > "All you have to do" is scan all the pages of the manuals, and
> > transfer the files of all the floppies...
>
> but,... i don't have anywhere to put it in cypher-space...
>
Without adding more discussion, to a very VERY long discussion:
There are many Web sites which host document archives of CP/M era
software and hardware. Ask those interested in YOUR documents, to
contribute costs or effort to photocopy, scan and to make PDF's.
Likewise, sites are archiving disk images - this is a recent and
repeated discussion in comp.os.cpm, look up those discussions for
specifics. As for starting a Web site and filling it with your work,
it is not difficult and will cost little or even zero - and it
provides a service to many people, some of whom will help you as well.
It is a means of archiving and distribution.
Herb Johnson
retrotechnology.com
Glen Herrmannsfeldt
> To use NFS in a useful way, and justify the implementation, CP/M had to
> get a BDOS extension that traps all calls dealing with a network drive.
> If I got things right, you would better run a CP/NET Server on Unix
> instead to save the implementation trouble and use UDP as transport.
> A minimal UDP implementation including ARP should only be a few pages
> of code and CP/NET is already there.
> If you simply want remote storage, then I suggest to use a storage
> access protocol, like NBD which originated in Linux, iSCSI or something
> like that.
The thought is to do it on the Spartan3E board. While one could do
a floppy interface, the board has ethernet hardware. Well, with
32MB of RAM, one could do a very large RAM disk, but external
storage would be nice. The idea was to do NFS and no disk interface.
> Microcomputers work most efficient if intelligent peripheral chips are
> used, and networking makes no exception here. A Z80 running CP/M would
> profit a lot from offloading the network stack to a network controller
> and such things exist already.
The offloading could be done in a variety of ways. With 32MB ram
and bank switch hardware, it could be done with 8080 code in
a different memory bank.
-- glen
Greegor
> On Fri, 5 Dec 2008 02:23:29 -0800 (PST), Greegor <Greego...@gmail.com>
Yeah, that makes perfect sense.
Sorry I got it wrong, but that one was from way back
in the cobwebs! LOL
Thanks again Allison!
Do you think that might have played out differently
had they done it for Z-80 rather than 8080?
I also vaguely remember that at least one
Ohio Scientific (Z-80) microcomputer also had
a 6502 processor and could execute code on
that, but the technical details on how to do this
were (at that time) not readily available.
The second processor was not merely
part of a subsystem, or so I was lead to believe.
Supposedly there were a few programs that
were able to tap into the 6502 for code
execution.
Was that also an overlooked capability?
Does this sort of fusion of dissimilar processors
fail to catch on simply because of cost factors?
Would dissimilar co-processors have any advantages
looking toward future computer systems?
Udo Munk
>> Message-ID: <E1L5BZL-...@stasis.kostecke.net>
>
> CBFalconer wrote:
>> Simply emulate a Hitachi 64180, which I believe is identical to a
>> Zilog Z180. It has a 1 megabyte address range, and is fully Z80
>> compatible.
>
> Talking of which, I've been adding '180 and '280 functionality to
> my Z80 emulator and I'd like to try banked CPM 3 on it. I spent a
> couple of hours last week roaming the web looking for a) CPM 3 and
> b) how to implement a suitable BIOS, but couldn't find anything.
> Any pointers?
>
> Ta.
http://www.unix4fun.org/z80pack/
CP/M 3 sources with banked BIOS for an emulated Z80 machine, should give
you everything you need to implement CP/M 3 on some sort of machine.
Enjoy,
Udo Munk
--
The real fun is building it and then using it...
no....@no.uce.bellatlantic.net
wrote:
>On Dec 5, 6:22 am, no.s...@no.uce.bellatlantic.net wrote:
The devil is in the details. The big part is the part always startup
in 8088 mode form reset.
I keep a Tandy 1000HX which came with V20 in it. interesting macine
from the clone era when not all clones were 100% IBM CP compatable.
Allison
Alec Voropay
> I just uploaded my Vector Graphic Z80 SBC design based on the Xilinx
> Spartan-3E Starter kit (US$149) towww.opencores.org:
>
> http://www.opencores.org/projects.cgi/web/vg_z80_sbc/overview
Did you see a FPGA clone of the russian Vector-06с home computer
(not a Vector Graphics clone):
http://en.wikipedia.org/wiki/Vector-06C
http://code.google.com/p/vector06cc/
software repository:
http://sensi.org/%7Esvo/scalar/
It emulates 8080 (as opencores T80 mode), graphic susbystem
on VGA with fine-tuned clocking, AY8912 sound and WD1793
controller. A second processor, a hidden 6502, used to read/write
floppy images on the conventional FAT-formatted SD card.
--
-=AV=-
Thierry B.
> GSX, which is a PORTABLE graphics system running under CP/M, CP/M-86,
> and MS-DOS. Find me a Linux graphics system running under CP/M...
May be, you can run cp/m programs under z80pack, and use
tek4014 capabilities of your xterm ?
--
"C makes it easy to shoot yourself in the foot. C++ makes it harder,
but when you do, it blows away your whole leg".
--{ Bjarne Stroustrup }--
Howard Harte
> On Dec 1, 5:16 am, Howard Harte <hhar...@yahoo.com> wrote:
>
> > I just uploaded my Vector Graphic Z80 SBC design based on the Xilinx
> > Spartan-3E Starter kit (US$149) towww.opencores.org:
>
> >http://www.opencores.org/projects.cgi/web/vg_z80_sbc/overview
>
> (not a Vector Graphics clone):
>
>
> software repository:
>
> http://sensi.org/%7Esvo/scalar/
>
> It emulates 8080 (as opencores T80 mode), graphic susbystem
> on VGA with fine-tuned clocking, AY8912 sound and WD1793
> controller. A second processor, a hidden 6502, used to read/write
> floppy images on the conventional FAT-formatted SD card.
>
> --
> -=AV=-
I hadn't seen that, thanks.
I added a Vector HD/FD disk controller to my design, using Flash for
storage. I was able to boot CP/M by using a Vector Graphic disk image
written to Flash. Since the DDR controller isn't working, I took it
out, and used block RAM from 0-2Fff and B000-FFFFh. That is enough
RAM for CP/M with a 12k TPA. Not much, but enough to run some
programs.
I found some bugs in the keyboard controller with shifted (caps)
characters and control characters not working. I also ran the
exz80all.Com instruction set test and found some bugs in the Z80
core. Pinpointing the exact instruction failures will be difficult,
but I think doable using SIMH as a reference.
I'll upload the latest files to opencores.org today.
-Howard
Udo Munk
> --{ Mr Emmanuel Roche, France a plopé ceci: }--
>
>> GSX, which is a PORTABLE graphics system running under CP/M, CP/M-86,
>> and MS-DOS. Find me a Linux graphics system running under CP/M...
>
> May be, you can run cp/m programs under z80pack, and use
> tek4014 capabilities of your xterm ?
You could if you can find the GSX Tek drivers anywhere, otherwise GSX is
working ok and the Tek emulation can be used from CP/M under z80pack, I
tested that.
Mr Emmanuel Roche, France
contains, among other things:
-------------
Programs that use graphics can be purchased from a number of
vendors. Three graphic programs from DRI are: DR GRAPH, DR DRAW, and
ACCESS 10.
DR GRAPH provides business graphics, including pie charts and
bar graphs; it can take data from many popular spreadsheet
programs and convert it to graphs, or data may be entered
directly from the keyboard.
DR DRAW is a very flexible free-form drawing and drafting
program; it even includes mouse support on the IBM PC.
ACCESS 10 is a Tektronix 4010/14 graphics terminal emulator that
allows a microcomputer to be connected to a graphics host
computer.
Other products using GSX include the SuperCalc 3 spreadsheet
from Sorcim, and GraphPlan from Chang Labs.
-----------
Unfortunately, I have never seen any mention of "Access-10" outside
this article. I someone out there knows more about it, or has a manual
of floppy...
Thierry B.
> Again, this shows your ignorance: the NEC uPD-7220 GDC could only
> manage 8 colors. We are talking of 25 years ago. How many screens
> would have displayed "good color depth" 25 years ago?
Around 1983, I was working for a french company, Numelec, who
build an image processing system, called Pericolor 1000. It was
build around a 8080 processor and can display (and process !)
pictures in true colors (3x8bits) in 256x256 on really nice
Barco monitors, or a standard Peritel TV set.
> If you "need to interact with a group of engineers on a project", use
> a network. Guess why Usenet was created? Why are you using it? In
> Black&White?
No, just using Subversion.
--
Sketch de Bigard:
Lorsqu'il voit dans la rue un plan avec marqué "Vous êtes ici" Il change de
trottoir et voit un autre plan avec marqué aussi "Vous êtes ici". Il se dit
alors : "Putain, on m'observe !" :-)
Tom Lake
> trottoir et voit un autre plan avec marqué aussi "Vous êtes ici". Il se dit
> alors : "Putain, on m'observe !" :-)
Yeah, how *do* those maps know exactly where I am at all times, anyway?
Tom Lake
Thierry B.
Offtopic and in french, but really funny...
http://la.buvette.org/vrac/gpsimac.ogg (1.5M)
--
Oups, I did it again, one more time...