26 Oct
2010
26 Oct
'10
11:55 p.m.
1. > The PDP-11 was in architectural ways more
important than the VAX, if
? > nothing else than just because the VAX was basically just extending the
? > PDP-11.
Just to throw another question into the fire:
Just how important was the PDP-11 or VAX-11/780 hardware architecture
in the grand scheme of things? Did either machine really bring
anything new to the table?
Don't confuse popularity with importance - we all know that sometimes
unpopular inventions are hugely influential (Alto?).
--
Will
27 Oct
27 Oct
9:04 a.m.
On 10/26/2010 11:55 PM, William Donzelli wrote:
Just how important was the PDP-11 or VAX-11/780
hardware architecture
in the grand scheme of things? Did either machine really bring
anything new to the table?
Don't confuse popularity with importance - we all know that sometimes
unpopular inventions are hugely influential (Alto?).
In this case I think popularity is important. The '11 instruction set
was somewhat of a
departure from other machines in that it was very "programmer
friendly". That, coupled
with the fact that it was very popular in universities led to a lot of
people embracing it.
The VAX was widely adopted in universities and at one point it could
have almost been
said that many people considered "all the worlds a VAX" since they were
so popular.
DEC was an alternative to IBM at the time; more accessible and more
oriented toward
individual scientific users.
(I don't know about you, but the 370 operators always got tense when I
touched
the console, but I would twist the power keys on most of the '11s and
VAXes I used
and I was alone in the room, and it was *much* quieter :-)
You could insert a small paragraph here about the role of unix and how
unix and
the pdp-11 and vax interacted. And then how the 680xx accelerated that
vector. It might be interesting to review how microcode (and nanocode)
affected all
three of those designs.
Both machines have been left in the dust by RISC architectures, mostly
because
of the notion that it's easier to accelerate (parallelize if you will)
MIPS style RISC
instruction sets. Even Intel turns ciscy X86 code into something which
looks
remarkably like MIPS code inside their chips.
Having said that, I would claim the '11 and VAX taught us that "going
more CISC"
was helpful but was ultimately doomed due to speed issues.
Much was learned from those architectures.
-brad
11:31 a.m.
In this case I think popularity is important. ?The
'11 instruction set was
somewhat of a
departure from other machines in that it was very "programmer friendly".
I think a lot of people would not agree with that. The PDP-10 people,
for example.
The VAX was widely adopted in universities and at one
point it could have
almost been
said that many people considered "all the worlds a VAX" since they were so
popular.
I am thinking more about the hardware development of the architecture
- were there any real innovations the DEC guys came up with and made
public when the machines were introduced?
(I don't know about you, but the 370 operators
always got tense when I
touched
the console,
Yes, when the EPO switch is pulled and it makes a sizable dent in the
company's stock, and maybe the Dow as a whole, well, yes, they tend to
be nervous.
You could insert a small paragraph here about the role
of unix and how unix
and
the pdp-11 and vax interacted.
Are you saying that the PDP-11 and VAX were the first machines where
the hardware and software were both considered and designed together?
Once again, I think a lot of people would not agree with that.
--
Will
28 Oct
28 Oct
1:12 p.m.
On Oct 27, 2010, at 11:31 AM, William Donzelli wrote:
agreed.
No, I didn't say that at all. I was just saying that the it would be interesting to
explore how the pdp-11 and VAX architectures influenced the design of unix,
and how unix, in turn, influenced software development as a whole. I thin
the two are interrelated.
As I recall, in 1983, running unix on a VAX was the "hot setup" for software
developers. Followed soon by running unix on a Sun...
-brad
Brad Parker
Heeltoe Consulting
781-483-3101
http://www.heeltoe.com
In this case I
think popularity is important. The '11 instruction set was
somewhat of a
departure from other machines in that it was very "programmer friendly".
I think a lot of people would not agree with that. The PDP-10 people,
for example. You could
insert a small paragraph here about the role of unix and how unix
and
the pdp-11 and vax interacted.
Are you saying that the PDP-11 and VAX were the first machines where
the hardware and software were both considered and designed together?
Once again, I think a lot of people would not agree with that.
1:20 p.m.
No, I didn't say that at all. ?I was just saying
that the it would be interesting to
explore how the pdp-11 and VAX architectures influenced the design of unix,
and how unix, in turn, influenced software development as a whole. ?I thin
the two are interrelated.
As I recall, in 1983, running unix on a VAX was the "hot setup" for software
developers. ?Followed soon by running unix on a Sun...
I was initially thinking about just the hardware architecture when the
machines were being designed, not the software aspect, nor what the
machines did (or influenced) after they were released. The
aforementioned memory mapped I/O, for example.
I do not think anyone can deny the importance of the PDP-11 and VAX to
the software development scene.
--
Will
1:51 p.m.
On 28 Oct 2010 at 13:20, William Donzelli wrote:
I was initially thinking about just the hardware
architecture when the
machines were being designed, not the software aspect, nor what the
machines did (or influenced) after they were released. The
aforementioned memory mapped I/O, for example.
Not to be too much of a wet blanket, but how many of those DEC-unique
innovations (even if you manage to assert that they originated with
DEC) persist in today's hardware? Do modern PCs use memory-mapped
I/O? The 68K, but for some Freescale relics, is history.
Major innovations, such as virtual memory and orthogonal instruction
sets and hardware-implemented stacks preceded the PDP-11.
Perhaps the major contribution of the PDP-11 and VAX was that they
were comparatively cheap for the processing power.
--Chuck
2:36 p.m.
On 10/28/10 1:51 PM, Chuck Guzis wrote:
Not to be too much of a wet blanket, but how many of
those DEC-unique
innovations (even if you manage to assert that they originated with
DEC) persist in today's hardware? Do modern PCs use memory-mapped
I/O? The 68K, but for some Freescale relics, is history.
Relics? The Freescale 68K variants are actually pretty mainstream.
-Dave
--
Dave McGuire
Port Charlotte, FL
4:14 p.m.
On 28 Oct 2010 at 14:36, Dave McGuire wrote:
Relics? The Freescale 68K variants are actually
pretty mainstream.
If Freescale survives, perhaps. But after being spun off from
Motorola and passed around as an LBO, Freescale has a lot of debt to
service.
If Freescale goes down the rathole, I suspect that 68K architecture
will disappear completely very quickly.
--Chuck
5:16 p.m.
On 10/28/10 4:14 PM, Chuck Guzis wrote:
Yes, but their market share is HUGE. Time will tell, but I for one
will be shocked out of my gourd if Freescale goes away.
Relics?
The Freescale 68K variants are actually pretty mainstream.
If Freescale survives, perhaps. But after being spun off from
Motorola and passed around as an LBO, Freescale has a lot of debt to
service. If Freescale goes down the rathole, I suspect that 68K
architecture
will disappear completely very quickly.
Well of course, as is the case with any (nearly-)single-source
architecture.
-Dave
--
Dave McGuire
Port Charlotte, FL
2:42 p.m.
On 2010 Oct 28, at 10:51 AM, Chuck Guzis wrote:
On 28 Oct 2010 at 13:20, William Donzelli wrote:
How about the notion that the PDP-11 was where several prior but
then-topical innovations coalesced into one machine/architecture?
The VAX was riding on the coattails of the market success of the -11
and provided existing PDP-11 installations with increasing demands with
a way forward, and it grew from there. That's a bit of a tautology, but
it is to say it did exactly what it was designed and marketed for.
On the topic of memory, I would agree with Johnny about "virtual
addresses", but differ on "virtual memory": virtual memory to me has
always meant demand-paging where the RAM address-space seen by the user
can be larger than the physical RAM, distinct from the simple mapping
of addresses, but the use of the term is a matter of definition.
I was initially thinking about just the hardware
architecture when the
machines were being designed, not the software aspect, nor what the
machines did (or influenced) after they were released. The
aforementioned memory mapped I/O, for example.
Not to be too much of a wet blanket, but how many of those DEC-unique
innovations (even if you manage to assert that they originated with
DEC) persist in today's hardware? Do modern PCs use memory-mapped
I/O? The 68K, but for some Freescale relics, is history.
Major innovations, such as virtual memory and orthogonal instruction
sets and hardware-implemented stacks preceded the PDP-11. Perhaps the major contribution of the PDP-11 and VAX
was that they
were comparatively cheap for the processing power.
.. the right product, at the right time, at the right price, from an
established and reliable manufacturer/source.
2:52 p.m.
Chuck Guzis [cclist at sydex.com] wrote:
Not to be too much of a wet blanket, but how many of
those DEC-unique
innovations (even if you manage to assert that they originated with
DEC) persist in today's hardware? Do modern PCs use memory-mapped
I/O? The 68K, but for some Freescale relics, is history.
Given that x86 is king today that means it all goes back to the 4004 :-)
Major innovations, such as virtual memory and
orthogonal instruction
sets and hardware-implemented stacks preceded the PDP-11.
Perhaps the major contribution of the PDP-11 and VAX was that they
were comparatively cheap for the processing power.
Neither VM nor a sensible ISA originated with VAX-11 (or PDP-11) but
these architectures were very popular and exposed many people to these
sorts of features.
Being hugely influential doesn't (necessarily) mean being the first with
anything.
VM, for example, is as old as the hills, I can find references as far
back as Atlas but I'd not be surprised if someone now points out a
known implementation even earlier.
BTW modern PCs are only just beginning to shake off their UARTs: wasn't
that an invention of one G. Bell for the PDP-11? (My "Computer
Engineering"
isn't to hand right now so I may be going out on a limb here ...)
Antonio
5:06 p.m.
arcarlini at iee.org wrote:
Given that x86 is king today that means it all goes
back to the 4004 :-)
Not really. The 8008 and 8080 architectures, and their successors, are
almost entirely unlike the 4004 and 4040. Intel worked with Busicom on
the definition of hte 4004 architecture, but the 8008 architecture came
from the Datapoint 2200.
3:34 p.m.
In article <4CC95597.9247.C408E0 at cclist.sydex.com>,
"Chuck Guzis" <cclist at sydex.com> writes:
Do modern PCs use memory-mapped I/O?
Just because a CPU architecture has IO instructions doesn't mean you
can't do memory-mapped I/O.
Memory mapped IO is ubuiqitous in MS-DOS based PCs, as it is generally
fastest to write directly to the memory-mapped display memory for CGA
and VGA displays. DOOM and many other MS-DOS VGA style games use this
technique to get the framerates they need.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
4:24 p.m.
On 28 Oct 2010 at 13:34, Richard wrote:
Just because a CPU architecture has IO instructions
doesn't mean you
can't do memory-mapped I/O.
Just because a CPU doesn't have any I/O instructions or I/O space,
doesn't mean that you can't invent them.
Memory mapped IO is ubuiqitous in MS-DOS based PCs, as
it is generally
fastest to write directly to the memory-mapped display memory for
CGA > and VGA displays.
And there were other PC displays, that didn't use memory-mapped video
at all. One liability with the PC was that you weren't able to move
display space anywhere--software assumed it to be at a fixed
location. That pretty much limited real-mode RAM space to 640K,
which eventually became something of a liability until protected-mode
CPUs with address remapping came along.
--Chuck
4:36 p.m.
In article <4CC979A0.22930.152134D at cclist.sydex.com>,
"Chuck Guzis" <cclist at sydex.com> writes:
You asked if it was used on modern PCs. The answer is yes. Every PC
these days supports standard VGA/monochrome display adapter
functionality which uses a memory-mapped framebuffer. Even the
high-end 3D cards support this functionality, albeit as a legacy from
past architectures.
Some modern video cards use memory-mapped IO for their interaction with
the CPU in terms of doing things like providing a common address space
for uploading textures and other 3D data. They don't provide
low-level control registers to the CPU for manipulating the video card
because that eliminates potential parallelism between the CPU and the
GPU. The GPU has become a smart coprocessor with shared address space
for transferring large amounts of data that needs to be directly
manipulated by the CPU and the GPU.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
Memory mapped
IO is ubuiqitous in MS-DOS based PCs, as it is generally
fastest to write directly to the memory-mapped display memory for
CGA > and VGA displays.
And there were other PC displays, that didn't use memory-mapped video
at all.
5:03 p.m.
On 28 Oct 2010 at 14:36, Richard wrote:
You asked if it was used on modern PCs. The answer is
yes. Every PC
these days supports standard VGA/monochrome display adapter
functionality which uses a memory-mapped framebuffer. Even the
high-end 3D cards support this functionality, albeit as a legacy from
past architectures.
I probably should have said "modern PCs exclusively use memory-mapped
I/O" a la PDP-11 or CP-1600 or MC6800...
Clearly there, the answer is "no". They have I/O address space and
use it.
Not that I think that memory-mapped I/O is a bad idea. Most
microcontrollers are witness to that.
--Chuck
5:40 p.m.
In article <4CC982AC.24989.1756BE2 at cclist.sydex.com>,
"Chuck Guzis" <cclist at sydex.com> writes:
I probably should have said "modern PCs
exclusively use memory-mapped
I/O" a la PDP-11 or CP-1600 or MC6800...
Well if you're just fishing for a self-fulfilling prophecy, I fail to
see the point of the question.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
28 Oct
28 Oct
4:56 p.m.
It was thus said that the Great Chuck Guzis once stated:
On 28 Oct 2010 at 13:34, Richard wrote:
The limit to 640k was for two reasons: 1) IBM decided to stick I/O and
ROM space starting at $A0000 and 2) MS-DOS required a contiguous memory
region (it couldn't deal with holes in memory).
Oh, and 3) nobody knows quite who said "640k should be enough for
everybody." It could have been IBM; it could have been Microsoft.
-spc (Gah! Why is this information still in my head? I haven't used
MS-DOS for fifteen years now!)
Just because a CPU architecture has IO
instructions doesn't mean you
can't do memory-mapped I/O.
Just because a CPU doesn't have any I/O instructions or I/O space,
doesn't mean that you can't invent them.
Memory mapped IO is ubuiqitous in MS-DOS based
PCs, as it is generally
fastest to write directly to the memory-mapped display memory for
CGA > and VGA displays.
And there were other PC displays, that didn't use memory-mapped video
at all. One liability with the PC was that you weren't able to move
display space anywhere--software assumed it to be at a fixed
location. That pretty much limited real-mode RAM space to 640K,
which eventually became something of a liability until protected-mode
CPUs with address remapping came along. 
29 Oct
29 Oct
9:51 a.m.
Sean Conner wrote:
Oh, and 3) nobody knows quite who said "640k
should be enough for
everybody." It could have been IBM; it could have been Microsoft.
I seem to remember there being some paraphrasing of this on some IBM marketing
material that TNMoC has on laserdisc - but it'd be easy enough to believe that
they meant "640K is enough for everybody in the context of today's industry"
rather than "640K is all that anyone will need in a computer, ever".
cheers
Jules
4:55 p.m.
> Just because a CPU architecture has IO
instructions doesn't mean you
> can't do memory-mapped I/O.
I beleive Steve Ciarcia said in one of the Circuit Cellar articles in
Byte many year ago that any processor that could access memory could have
memory-mapped I/O.
I would add that in the case of a system (rather than a bare processor),
it helps if there's some spare space in the memroy map :-).
And there were other PC displays, that didn't use memory-mapped video
if by 'PC' here you mean something descended from the IBM5150 then I am
curious. All the IBM dipslay adapdaters that I can think of were memory
mapped. Yes, even the PGC had a memory-mapped CGA emulation.
I do rememebr reading about TMS34010 development boards for the PC. Is
that the sort of thing you were thinking of?
-tony
30 Oct
30 Oct
1:31 a.m.
On 10/29/10 4:55 PM, Tony Duell wrote:
He certainly did, I remember that article well. He started the
article by relating a conversation that he'd had with a fellow computer
guy, and they were arguing (I think) 6800 vs. Z80. He said something
like "When he said the 6800 had memory-mapped I/O and the Z80 didn't, I
knew he didn't know what it was". :)
-Dave
--
Dave McGuire
Port Charlotte, FL
> Just
because a CPU architecture has IO instructions doesn't mean you
> can't do memory-mapped I/O.
I beleive Steve Ciarcia said in one of the Circuit Cellar articles in
Byte many year ago that any processor that could access memory could have
memory-mapped I/O. 
28 Oct
28 Oct
7:05 p.m.
On Thu, 28 Oct 2010, Chuck Guzis wrote:
Perhaps the major contribution of the PDP-11 and VAX
was that they
were comparatively cheap for the processing power.
"They DEFINED the MINI-computer market segment!"
There were years when fresh CS graduates didn't accept the existence of
anything else - much like Windoze now is for the general public.
--
Grumpy Ol' Fred cisin at xenosoft.com
2:05 p.m.
From: Brad Parker
Sent: Thursday, October 28, 2010 10:12 AM
On Oct 27, 2010, at 11:31 AM, William Donzelli wrote:
>> You could insert a small paragraph here about
the role of unix and how
>> unix and the pdp-11 and vax interacted.
> Are you saying that the PDP-11 and VAX were the
first machines where
> the hardware and software were both considered and designed together?
> Once again, I think a lot of people would not agree with that.
No, I didn't say that at all. I was just saying
that the it would be
interesting to explore how the pdp-11 and VAX architectures influenced
the design of unix, and how unix, in turn, influenced software
development as a whole. I thin the two are interrelated.
From what Thompson and Ritchie have written about the
origins of Unix,
the PDP-11's architecture had very little to do with how it was
created.
The original PDP-7 Unix (specific I/O ports addressed in I/O instructions,
18-bit words, word addressing) was taken up and ported to the PDP-11 with
few or no user-visible changes.
Because the VAX offered a virtual memory capability (hi, Johnny! :-), it
did change the way Unix developed, but so did other ports (Interdata, for
example, and even the IBM Series/1).
As I recall, in 1983, running unix on a VAX was the
"hot setup" for
software developers. Followed soon by running unix on a Sun...
You seem to forget other mini manufacturers. The University of Chicago,
for example, eschewed the VAX altogether in favour of a Pyramid 90x in
1983. In fact, when I took classes in Unix administration in the late
c. 1990 while working at Stanford, the outfit teaching the classes used a
Pyramid system so that we could learn both BSD and ATT style admin tools.
(Stanford is of course where SUN came into being.)
Rich Alderson
Vintage Computing Sr. Server Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.PDPplanet.org/
http://www.LivingComputerMuseum.org/
11:59 p.m.
As I recall, in 1983, running unix on a VAX was the
"hot setup" for software
developers. Followed soon by running unix on a Sun...
The first reference I remember to a VAX as a kid was reading an issue of
ENTER magazine (I doubt anyone here remembers it) about the development of
Williams' Sinistar and this mysterious huge computer they did crossassembly
on (for the 6809).
--
------------------------------------ personal: http://www.cameronkaiser.com/ --
Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckaiser at floodgap.com
-- You can't kill me because I've got magic aaargh. -- Terry Pratchett --------
29 Oct
29 Oct
9:40 a.m.
On Thu, 28 Oct 2010, Cameron Kaiser wrote:
This? http://www.bombjack.org/commodore/magazines/enter/enter.htm
Enjoy. ;)
g.
--
Proud owner of F-15C 80-0007
http://www.f15sim.com - The only one of its kind.
http://www.simpits.org/geneb - The Me-109F/X Project
ScarletDME - The red hot Data Management Environment
A Multi-Value database for the masses, not the classes.
http://www.scarletdme.org - Get it _today_!
Political correctness is a doctrine, fostered by a delusional, illogical
minority, and rabidly promoted by an unscrupulous mainstream media, which
holds forth the proposition that it is entirely possible to pick up a turd
by the clean end.
As I recall,
in 1983, running unix on a VAX was the "hot setup" for software
developers. Followed soon by running unix on a Sun...
The first reference I remember to a VAX as a kid was reading an issue of
ENTER magazine (I doubt anyone here remembers it) about the development of
Williams' Sinistar and this mysterious huge computer they did crossassembly
on (for the 6809).
11:07 a.m.
New subject: ENTER magazine was Re: Happy Birthday VAX 11/780
The first
reference I remember to a VAX as a kid was reading an issue of
ENTER magazine (I doubt anyone here remembers it) about the development of
Williams' Sinistar and this mysterious huge computer they did crossassembly
on (for the 6809).
This? http://www.bombjack.org/commodore/magazines/enter/enter.htm
4964
days inactive
4967
days old
28 comments
15 participants
participants (15)
-
arcarlini＠iee.org -
ard＠p850ug1.demon.co.uk -
brad＠heeltoe.com -
cclist＠sydex.com -
cisin＠xenosoft.com -
eric＠brouhaha.com -
geneb＠deltasoft.com -
hilpert＠cs.ubc.ca -
jules.richardson99＠gmail.com -
legalize＠xmission.com -
mcguire＠neurotica.com -
RichA＠vulcan.com -
spc＠conman.org -
spectre＠floodgap.com -
wdonzelli＠gmail.com