Many hundred billion years ago, there existed a bizzare world that most 
of us have only read about in books. Plastics did not yet exist, so 
small children's toys were made of wood or metal, because there was 
literally nothing cheaper. Wires were insulated with fabic. If you 
wanted to make a telephone call, you did not *dial* the telephone 
number; you told the human being at the other end of the line which 
number you wanted to be connected to, and they would physically plug in 
the right cable.

Somewhere in this primordial soup, the first computers somehow came into 
being. According to the dusty history books, the very first computers 
used technologies such as electro-mechanical relays, vacuum tubes, drum 
memory, punch cards and so forth. And, in the beginning, all of the 
components were wired together by hand. This resulted in huge pieces of 
engineering which filled entire buildings, consumed insane quantities of 
electricity, requires specialist cooling systems, and had _vastly_ less 
computing power than the microcontroller in your washing machine.

Was anybody actually there? Did any of you guys see this happen? What 
was it like? And did mainframes really come in bright orange cases with 
inch-square glowy buttons on the front that randomly twinkle?

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

On 5/17/2010 2:54 PM, Orchid XP v8 wrote:

>
> Was anybody actually there? Did any of you guys see this happen? What
> was it like? And did mainframes really come in bright orange cases with
> inch-square glowy buttons on the front that randomly twinkle?
>

;) I remember as a kid in the 80's walking into an office that had a 
computer room. A real-life room-sized mainframe. One wall had a bank of 
tape machines, the CPU was a box the size of a Smart car sort of in the 
middle of the room, and there was a huge array of disks that took up the 
rest of the room.

All controlled by a tiny terminal in the corner flanked by an 8" floppy 
drive, and a few other binders and such.

It was quite neat to see!

-- 
~Mike

Post a reply to this message

Orchid XP v8 <voi### [at] devnull> wrote:
> Many hundred billion years ago, there existed a bizzare world that most
> of us have only read about in books. Plastics did not yet exist, so
> small children's toys were made of wood or metal, because there was
> literally nothing cheaper. Wires were insulated with fabic. If you
> wanted to make a telephone call, you did not *dial* the telephone
> number; you told the human being at the other end of the line which
> number you wanted to be connected to, and they would physically plug in
> the right cable.
>
> Somewhere in this primordial soup, the first computers somehow came into
> being. According to the dusty history books, the very first computers
> used technologies such as electro-mechanical relays, vacuum tubes, drum
> memory, punch cards and so forth. And, in the beginning, all of the
> components were wired together by hand. This resulted in huge pieces of
> engineering which filled entire buildings, consumed insane quantities of
> electricity, requires specialist cooling systems, and had _vastly_ less
> computing power than the microcontroller in your washing machine.
>
> Was anybody actually there? Did any of you guys see this happen? What
> was it like? And did mainframes really come in bright orange cases with
> inch-square glowy buttons on the front that randomly twinkle?
>
> --
> http://blog.orphi.me.uk/
> http://www.zazzle.com/MathematicalOrchid*

This is an image of the first mainframe I worked with in 1969.

http://www-03.ibm.com/ibm/history/exhibits/mainframe/images/2423PH2050.jpg

It was an IBM 360 model 50. The tape drives in the background are IBM 2400
models.  The machine was controlled from the IBM 1052 console, a modified IBM
typewriter.  The IBM 2501 just in front and to the left of the operator could
read about 1000 cards per minute.  The IBM 2311 disk drives in the foreground
had removable disk packs which held a little more than 7 MB IIRC. As you can see
there were quite a few flashing lights on the front panel. Besides showing the
status info, you could use them to read/alter the contents of memory.
Sometime I think they were a lot more fun then.

Isaac.

Post a reply to this message

Clarence1898 wrote:

> This is an image of the first mainframe I worked with in 1969.
> 
> http://www-03.ibm.com/ibm/history/exhibits/mainframe/images/2423PH2050.jpg

Trippy! :-D

> It was an IBM 360 model 50. The tape drives in the background are IBM 2400
> models.  The machine was controlled from the IBM 1052 console, a modified IBM
> typewriter.  The IBM 2501 just in front and to the left of the operator could
> read about 1000 cards per minute.  The IBM 2311 disk drives in the foreground
> had removable disk packs which held a little more than 7 MB IIRC.

Any idea how much all this gear cost?

> As you can see
> there were quite a few flashing lights on the front panel. Besides showing the
> status info, you could use them to read/alter the contents of memory.
> Sometime I think they were a lot more fun then.

Presumably not at the times when you wanted to get some actual *work* 
done? ;-)

Post a reply to this message

Mike Raiford wrote:

> ;) I remember as a kid in the 80's walking into an office that had a 
> computer room. A real-life room-sized mainframe. One wall had a bank of 
> tape machines, the CPU was a box the size of a Smart car sort of in the 
> middle of the room, and there was a huge array of disks that took up the 
> rest of the room.
> 
> All controlled by a tiny terminal in the corner flanked by an 8" floppy 
> drive, and a few other binders and such.
> 
> It was quite neat to see!

I guess it's like the steam engines of old; today we have much more 
powerful machines that take up a fraction of the power and space. But 
still, few things are more impressive than watching a 50T steam engine 
in motion...

Post a reply to this message

I had always assumed that the first computers were like current 
computers, just using relays or whatever instead of transisters, and 
with vastly inferior specifications.

However, it appears that this isn't the case.

For example, I thought they all used latch circuits for memory, but 
apparently not. There were things like core memory, which I'd never 
heard of. Presumably it's faster and cheaper to make core memory as 
opposed to wiring up thousands of latch circuits?

Another example. According to legend, there was a time when if you 
wanted to run a program, you used a machine not unlike a typewriter to 
punch holes into a card. You "type in" the program onto punch cards like 
this, and only once the entire program and all its data has been punched 
do you even go near the actual *computer*. You feed the cards into a 
reader. It reads them all, and then spends the next six months running 
the program. Finally, you get a stack of new punched cards representing 
the results.

Does anybody know approximately when this time was?

For that matter, does anybody have a broad timeline of when various 
technologies were in use? What are the dates for things like core 
memory, drum memory, punch cards, magnetic tape, relays, vacuum tubes, 
transistors, ICs, etc?

Was there ever a time when programs were entered into memory via 
switches rather than some other medium?

Was there ever a "punched tape" medium similar to punch cards?

Similarly, you hear people talk about the VAX, the PDP, the varouis IBM 
mainframes and Cray supercomputers. Does anybody know the timeline for 
these, the technologies used and the basic design and performance details?

(Sure, you can look up individual questions on Wikipedia, but the 
articles tend to contain huge amounts of minute detail about specific 
things. I'm trying to get a general overview of an entire era.)

Post a reply to this message

> Was there ever a time when programs were entered into memory via 
> switches rather than some other medium?

I do that all the time :-)

Post a reply to this message

On 18/05/2010 11:56 AM, Invisible wrote:
> Was there ever a time when programs were entered into memory via
> switches rather than some other medium?
>
> Was there ever a "punched tape" medium similar to punch cards?

When I started working in computing in 1969, the Honeywell H416 then 
later H316 was booted by entering about a dozen binary codes into the 
switch register. This told the CPU which peripheral to read and where in 
memory to but the loader program, which was read using a paper tape 
reader. After the loader tape was read you then read the program using 
the paper tape reader. Only then could you enter any data. In 1976 
Burroughs Machines were manufacturing 96 column card readers.

The first computer I ever saw was in 1997 at Glasgow University. It was 
an analogue machine that used valves.


-- 

Best Regards,
	Stephen

Post a reply to this message

>> Was there ever a time when programs were entered into memory via 
>> switches rather than some other medium?
> 
> I do that all the time :-)

*sigh*

I ment where you have a bank of switches to select a memory address, and 
a second bank of switches to select what datum to write there.

(Or even just a row of switches for every memory register - but I doubt 
anybody ever did _that_...)

Post a reply to this message

Stephen wrote:

> When I started working in computing in 1969, the Honeywell H416 then 
> later H316

Oh yes - apparently Honeywell used to make computers...

> was booted by entering about a dozen binary codes into the 
> switch register. This told the CPU which peripheral to read and where in 
> memory to but the loader program, which was read using a paper tape 
> reader.

Does the tape have a single row of holes? Or does it have multiple 
parallel rows? How easy is it to tear the stuff?

> After the loader tape was read you then read the program using 
> the paper tape reader. Only then could you enter any data.

 From tape or card, presumably?

> In 1976 Burroughs Machines were manufacturing 96 column card readers.

I gather Burroughs got bought to form Unisys?

> The first computer I ever saw was in 1997 at Glasgow University. It was 
> an analogue machine that used valves.

1997? o_O

Post a reply to this message

>>> Was there ever a time when programs were entered into memory via 
>>> switches rather than some other medium?
>>
>> I do that all the time :-)
>
> *sigh*

:-D

> I ment where you have a bank of switches to select a memory address, and a 
> second bank of switches to select what datum to write there.

I did that at university.  We had a hex keypad to write machine code 
instructions to RAM.  IIRC we had to write a program to do some thermostat 
control of a heater or something - extra marks for proportional control. 
But then this was just part of the CPU course to get us familiar with 
instructions and stuff.

Post a reply to this message

Invisible <voi### [at] devnull> wrote:
> Clarence1898 wrote:
>
> > This is an image of the first mainframe I worked with in 1969.
> >
> > http://www-03.ibm.com/ibm/history/exhibits/mainframe/images/2423PH2050.jpg
>
> Trippy! :-D
>
> > It was an IBM 360 model 50. The tape drives in the background are IBM 2400
> > models.  The machine was controlled from the IBM 1052 console, a modified IBM
> > typewriter.  The IBM 2501 just in front and to the left of the operator could
> > read about 1000 cards per minute.  The IBM 2311 disk drives in the foreground
> > had removable disk packs which held a little more than 7 MB IIRC.
>
> Any idea how much all this gear cost?

Depending on the size of the original 360 series machines, purchase price ranged
from around $100000 to several million.  The companies I worked for always
leased their machines.  I wasn't involved in the financial end, but I think we
paid somewhere around $20000 to $30000 per month for the 360/50.
That was for a machine with 128KB of ram, single processor rated at about 0.2
MIPS, with 16 IBM2314 disk drives with 29MB each, and 8 7-track tape drives, an
IBM 2540 card read/punch, plus a couple of IBM 2701 telecommunications
controllers.

>
> > As you can see
> > there were quite a few flashing lights on the front panel. Besides showing the
> > status info, you could use them to read/alter the contents of memory.
> > Sometime I think they were a lot more fun then.
>
> Presumably not at the times when you wanted to get some actual *work*
> done? ;-)

The machine was busy running production work weekdays and saturday. Since I
installed the os, IBM OS/360, all my testing had to be done on Sunday.  So a lot
of desk checking and setup through the week getting ready for the weekend.

Isaac.

Post a reply to this message

>> I ment where you have a bank of switches to select a memory address, 
>> and a second bank of switches to select what datum to write there.
> 
> I did that at university.  We had a hex keypad to write machine code 
> instructions to RAM.  IIRC we had to write a program to do some 
> thermostat control of a heater or something - extra marks for 
> proportional control. But then this was just part of the CPU course to 
> get us familiar with instructions and stuff.

Right. So you type stuff in hex rather than raw binary, and it 
automatically advances to the next address?

Post a reply to this message

>> I did that at university.  We had a hex keypad to write machine code 
>> instructions to RAM.  IIRC we had to write a program to do some 
>> thermostat control of a heater or something - extra marks for 
>> proportional control. But then this was just part of the CPU course to 
>> get us familiar with instructions and stuff.
>
> Right. So you type stuff in hex rather than raw binary, and it 
> automatically advances to the next address?

Yeh something like that, I don't remember the exact details.  I think you 
could enter a memory address and hit the "address" button or something and 
it would then accept instructions from that point, and then there was a 
"run" button.  Pretty basic, but armed with the instruction set datasheet 
you could actually write stuff that worked.  IIRC they had a few channels of 
analogue and digital IO mapped to certain addresses that were connected to 
useful stuff like LEDs, temperature sensors and heaters.

Post a reply to this message

>> Right. So you type stuff in hex rather than raw binary, and it 
>> automatically advances to the next address?
> 
> Yeh something like that, I don't remember the exact details.  I think 
> you could enter a memory address and hit the "address" button or 
> something and it would then accept instructions from that point, and 
> then there was a "run" button.  Pretty basic, but armed with the 
> instruction set datasheet you could actually write stuff that worked.  
> IIRC they had a few channels of analogue and digital IO mapped to 
> certain addresses that were connected to useful stuff like LEDs, 
> temperature sensors and heaters.

Mmm, tasty.

The closest I ever came to doing something like this was assembling a 
program for my dad's C64 by hand. (I still have the book with the 
op-code table in it. And, 20 years later, I still remember that RTS is 
96 decimal. How sad is that?)

Post a reply to this message

On 18/05/2010 12:36 PM, Invisible wrote:
> Stephen wrote:
>
>> When I started working in computing in 1969, the Honeywell H416 then
>> later H316
>
> Oh yes - apparently Honeywell used to make computers...

Yes the ones I worked on were for process control and they had core 
memory. Ferrite rings threaded with address wires and read write wires, 
by hand. The H316 could have 4 K of memory.


> Does the tape have a single row of holes? Or does it have multiple
> parallel rows? How easy is it to tear the stuff?
>

Parallel rows, one traction hole at twice the pitch of the data holes 
and seven or eight data holes depending on the parity.

>> After the loader tape was read you then read the program using the
>> paper tape reader. Only then could you enter any data.
>
>  From tape or card, presumably?

Yes or from sense switches from the process it controlled.

>
>> In 1976 Burroughs Machines were manufacturing 96 column card readers.
>
> I gather Burroughs got bought to form Unisys?
>

After  my time.

>> The first computer I ever saw was in 1997 at Glasgow University. It
>> was an analogue machine that used valves.
>
> 1997? o_O

Oops! 1967 :-)

-- 

Best Regards,
	Stephen

Post a reply to this message

http://www.columbia.edu/acis/history/fisk.pdf

Hmm, interesting. Some of the guys I went to uni with learned COBOL. All 
they could tell me about it is "if you miss out a dot, you get an error 
message 300 lines later". Sounds delightful...

Post a reply to this message

Invisible <voi### [at] devnull> wrote:
> I had always assumed that the first computers were like current
> computers, just using relays or whatever instead of transisters, and
> with vastly inferior specifications.
>
> However, it appears that this isn't the case.
>
> For example, I thought they all used latch circuits for memory, but
> apparently not. There were things like core memory, which I'd never
> heard of. Presumably it's faster and cheaper to make core memory as
> opposed to wiring up thousands of latch circuits?
>

"Core" memory was invented in the late 40s, but didn't come into wide use until
the early 50s.  I think the IBM 360s used core memory. The IBM 370s and all
later machines used integrated circuits for memory.  We added 1MB of memory to
an IBM 370/168 in the mid 70s.  Price? $50000. There were several advantages to
core memory.  First its not volatile, you don't loose the contents of memory
when you power off the system.  Second its relatively immune to electromagnetic
radiation, so it does not require shielding when used in hostile environements.
Many military aircraft used core memory in their avionics.  I think the space
shuttles originally used core memory in their computers.

> Another example. According to legend, there was a time when if you
> wanted to run a program, you used a machine not unlike a typewriter to
> punch holes into a card. You "type in" the program onto punch cards like
> this, and only once the entire program and all its data has been punched
> do you even go near the actual *computer*. You feed the cards into a
> reader. It reads them all, and then spends the next six months running
> the program. Finally, you get a stack of new punched cards representing
> the results.
>
> Does anybody know approximately when this time was?

When I was at university in the mid to late 60s, I took a FORTRAN programming
course.  Punch program into cards, place in card tray, come back 4 to 6 hours
later and pick up your card deck and output listing.  Fix your typos and bugs
and repeat the above process until the program works.  At my first commercial
job, it was similar.  Punch you program into cards, place in the card reader,
and wait for the output to appear on the printer. Where I worked cards were used
until the mid 1980s.  We had cabinet after cabinet filled with punched card
programs.  The fun began after you punched in your 1000 card program and dropped
the deck on the way to the card reader.  After you did that a couple of times,
you started adding sequence numbers to the deck.

>
> For that matter, does anybody have a broad timeline of when various
> technologies were in use? What are the dates for things like core
> memory, drum memory, punch cards, magnetic tape, relays, vacuum tubes,
> transistors, ICs, etc?
>
> Was there ever a time when programs were entered into memory via
> switches rather than some other medium?
>
> Was there ever a "punched tape" medium similar to punch cards?
>
Yes. Several different varieties of tape. Different widths, different hole
styles etc.  I once had to write a driver for the IBM 1056 paper tape reader on
an IBM 370/158 because IBM did not support that device on that machine.  That
was kind of fun to do. That was in the early 70s?

> Similarly, you hear people talk about the VAX, the PDP, the varouis IBM
> mainframes and Cray supercomputers. Does anybody know the timeline for
> these, the technologies used and the basic design and performance details?
>
> (Sure, you can look up individual questions on Wikipedia, but the
> articles tend to contain huge amounts of minute detail about specific
> things. I'm trying to get a general overview of an entire era.)

I think you will have to look it up.  There is just too much information to
enter here.  I have work exclusively on IBM machines my entire career.  So I
can't say anything about other manufacturers. Very briefly IBM 360 series,
transistor circuits, core memory, mid 60s to early 70s. Early IBM 370 series in
the early 70s, integrated circuits, to later 370 series with higher density
circuits in the mid 80s.  The 3090 series machines with very high density
circuits until the mid 90s.  The s/390 series user cmos technology and were air
cooled.  The older machines used bipolar technology and many were water cooled.
Finally the z/series, again cmos technology.  Speeds ranged from .2 MIPS for a
360/50, 1 MIP for a 370/158, 15 MIPS for a 3083. The current machines I work
with today, 2 z/9 BC processors are rated at 600 and 335 MIPS.
The top of the line z/10 is almost 30000 MIPS.

Isaac

Post a reply to this message

> "Core" memory was invented in the late 40s, but didn't come into wide use until
> the early 50s.  I think the IBM 360s used core memory.

How about drum memory? Or the various delay lines?

> The IBM 370s and all later machines used integrated circuits for memory.

As in, whole memory chunks on a single IC? Or just two latches per IC? 
Or...?

> We added 1MB of memory to an IBM 370/168 in the mid 70s.  Price? $50000.

Ouch! o_O

It still somewhat blows my mind that you could do anything useful with 
so little memory. Presumably for processing large datasets, most of the 
data at any one time would be in secondary storage?

> There were several advantages to core memory.

Yes, so I see.

> When I was at university in the mid to late 60s, I took a FORTRAN programming
> course.  Punch program into cards, place in card tray, come back 4 to 6 hours
> later and pick up your card deck and output listing.  Fix your typos and bugs
> and repeat the above process until the program works.

Heh, sounds line fun. According to the account I just read, you actually 
put one statement per punch card. Is that right? I always assumed that 
each card just held X characters of data, and you type until the card is 
full, then move to the next one. I didn't realise there was an actual 
"significance" to card bounderies. Or is that just for the benefit of 
the humans?

> Where I worked cards were used until the mid 1980s.

Wow. I had no idea it persisted that long. In the mid 1980s, I had a C64 
to play with, which stored data on audio tape - for more data than any 
stack of punch cards. And I don't think the C64 was even particularly 
expensive... (I know of a few people who actually used the C64 for work 
purposes, would you believe.)

> The fun began after you punched in your 1000 card program and dropped
> the deck on the way to the card reader.  After you did that a couple of times,
> you started adding sequence numbers to the deck.

Hell yeah! ;-)

>> Was there ever a "punched tape" medium similar to punch cards?
>
> Yes. Several different varieties of tape.

OK. I guess each manufacturer probably had their own style.

> I once had to write a driver for the IBM 1056 paper tape reader on
> an IBM 370/158 because IBM did not support that device on that machine.  That
> was kind of fun to do. That was in the early 70s?

Hmm, OK.

What sort of access speeds do you get for reading or writing punched 
card or tape?

>> Similarly, you hear people talk about the VAX, the PDP, the varouis IBM
>> mainframes and Cray supercomputers. Does anybody know the timeline for
>> these, the technologies used and the basic design and performance details?
> 
> I think you will have to look it up.  There is just too much information to
> enter here.  I have work exclusively on IBM machines my entire career.  So I
> can't say anything about other manufacturers.

Fair enough.

> Very briefly IBM 360 series, transistor circuits, core memory,
> mid 60s to early 70s.

As in, discrete transistors on PCBs?

Any ideas on typical memory capacity / clock speed? Data path widths?

> Early IBM 370 series in
> the early 70s, integrated circuits, to later 370 series with higher density
> circuits in the mid 80s.

Are we talking ICs with single logic gates, small logic blocks, or an 
entire CPU on a chip?

> The 3090 series machines with very high density
> circuits until the mid 90s.

I had no idea people were still making mainframes in the 90s.

> The s/390 series user cmos technology and were air
> cooled.  The older machines used bipolar technology and many were water cooled.
> Finally the z/series, again cmos technology.

Nice consistent naming scheme. Heh. :-/

> Speeds ranged from .2 MIPS for a
> 360/50, 1 MIP for a 370/158, 15 MIPS for a 3083. The current machines I work
> with today, 2 z/9 BC processors are rated at 600 and 335 MIPS.
> The top of the line z/10 is almost 30000 MIPS.

Now if I could figure out what a typical MIPS rating for a normal PC 
today is, I'd have something dissimilar to compare it to. :-}

Post a reply to this message

On 18/05/2010 3:14 PM, Invisible wrote:
> What sort of access speeds do you get for reading or writing punched
> card or tape?

Some of the fast paper tape readers would stream the tape about 15 feet 
through the air before it was collected in a large bin. The fun part was 
rewinding into a roll.

BTW I forgot. Plain paper tape was easily torn but the tapes used in the 
fast readers had a layer of Mylar so they could not be torn, the paper 
backing could separate though.


-- 

Best Regards,
	Stephen

Post a reply to this message

>> We added 1MB of memory to an IBM 370/168 in the mid 70s.  Price? $50000.
>
> Ouch! o_O
>
> It still somewhat blows my mind that you could do anything useful with so 
> little memory.

You could do useful stuff with *far* less.

> Wow. I had no idea it persisted that long. In the mid 1980s, I had a C64 
> to play with, which stored data on audio tape - for more data than any 
> stack of punch cards. And I don't think the C64 was even particularly 
> expensive... (I know of a few people who actually used the C64 for work 
> purposes, would you believe.)

I can remember when I was younger (mid 80s) my dad writing a program to 
simulate the 2D flow of cylindrical shaped objects into different shaped 
channels (with graphics) - that was with 32K of RAM.  Useful work (even if 
it took about 10 seconds to eventually figure out where each one went!), and 
I'm sure there are loads more examples of doing useful stuff with just a few 
KB of RAM.

Post a reply to this message

On Mon, 17 May 2010 16:08:29 -0500, Mike Raiford wrote:

> On 5/17/2010 2:54 PM, Orchid XP v8 wrote:
> 
> 
>> Was anybody actually there? Did any of you guys see this happen? What
>> was it like? And did mainframes really come in bright orange cases with
>> inch-square glowy buttons on the front that randomly twinkle?
>>
>>
> ;) I remember as a kid in the 80's walking into an office that had a
> computer room. A real-life room-sized mainframe. One wall had a bank of
> tape machines, the CPU was a box the size of a Smart car sort of in the
> middle of the room, and there was a huge array of disks that took up the
> rest of the room.
> 
> All controlled by a tiny terminal in the corner flanked by an 8" floppy
> drive, and a few other binders and such.
> 
> It was quite neat to see!

My university had an IBM 4361 mainframe (IIRC), and I can remember in 6th 
grade going to a class where I got to play in the server room with some 
PDP mainframes (I helped with the backups one day).

My first job out of college had a fairly large AS/400 system (D40?  I 
think) and an IBM System/34.

The S/34 didn't have very many lights on it, but the AS/400 lit up light 
a Christmas tree when we turned the lights off in the bunker that was our 
computer room.

Jim

Post a reply to this message

Invisible <voi### [at] devnull> wrote:

> How about drum memory? Or the various delay lines?
>

Don't know, was before my time.

> As in, whole memory chunks on a single IC? Or just two latches per IC?
> Or...?

I don't know very much about the actual electronics. You will have to go to
google or wikipedia for that info.  I remember the 3081 had what were called
TCMs.  Fully encased, water cooled modules with about a thousand pins on the
back.  The CEs had lots of fun replacing them without bending a pin in the
process.

> It still somewhat blows my mind that you could do anything useful with
> so little memory. Presumably for processing large datasets, most of the
> data at any one time would be in secondary storage?

Most really large files were on tape, you just read the data serially.
For highly accessed files, or files that needed random access were kept on disk.

> Heh, sounds line fun. According to the account I just read, you actually
> put one statement per punch card. Is that right? I always assumed that
> each card just held X characters of data, and you type until the card is
> full, then move to the next one. I didn't realise there was an actual
> "significance" to card bounderies. Or is that just for the benefit of
> the humans?

The file system for IBM mainframes are record oriented, not stream.
Files are created with a defined record format and record length.
There is no end of record indicator such as CR or LF as in PC files.
Since cards were 80 columns long, many source files today are also 80 byte
records.  The norm was for program source files was column 1 to 72 was data
and column 73 to 80 was sequence numbers. For most COBOL programs the sequence
numbers were in cols 1 to 6.

> What sort of access speeds do you get for reading or writing punched
> card or tape?

As I recall about 60 characters per second.  The tape was paper, was 8 holes
wide, and easily broken or scrunched.

> As in, discrete transistors on PCBs?
>

Again, I don't know much about the electronic packaging.

> Any ideas on typical memory capacity / clock speed? Data path widths?

On some of the older machines I worked with:
  360/50 128KB
  370/158 1MB
  370/168 3MB
  3033U 4MB
  3081Q 8MB
On the latest machine:
  z/9 BC S03 - 16GB

> Nice consistent naming scheme. Heh. :-/

You can always depend on IBM to be consistent.

> Now if I could figure out what a typical MIPS rating for a normal PC
> today is, I'd have something dissimilar to compare it to. :-}

It is rather difficult to compare since they have such different architectures.
I doubt that the internal processor speed is that much different than current
PCs, but the i/o bandwidth is much higher in the mainframe.
For example our two machines, which are on the small side for mainframes, are
connected to a disk storage unit containing 15TB of data.  There are 4 high
speed fiber optic channels from each processor to this unit. We can easily
sustain i/o rates of over 10000 i/o operations per second, with peaks in the
20000 to 40000 range.  It is very rare for any single job to saturate the cpu.
By running many jobs concurrently we make full use of the machine.  If I could
port POVRAY to our mainframe, I'm sure it would be slower than most current PCs.
 But cpu intensive work is not what the mainframe does best.

Isaac

Post a reply to this message

Orchid XP v8 <voi### [at] devnull> wrote:
> Many hundred billion years ago, there existed a bizzare world that most 
> of us have only read about in books. Plastics did not yet exist, so 
> small children's toys were made of wood or metal, because there was 
> literally nothing cheaper.

  There were no children many hundred billion years ago.

-- 
                                                          - Warp

Post a reply to this message

Warp wrote:
> Orchid XP v8 <voi### [at] devnull> wrote:
>> Many hundred billion years ago, there existed a bizzare world that most 
>> of us have only read about in books. Plastics did not yet exist, so 
>> small children's toys were made of wood or metal, because there was 
>> literally nothing cheaper.
> 
>   There were no children many hundred billion years ago.

I was waiting for that one... *sigh*

Post a reply to this message

>> How about drum memory? Or the various delay lines?
> 
> Don't know, was before my time.

OK. So care is newer than either of those.

> I don't know very much about the actual electronics.

OK.

> Most really large files were on tape, you just read the data serially.
> For highly accessed files, or files that needed random access were kept on disk.

So the concept of a filesystem storing named files already existed at 
this time?

> The file system for IBM mainframes are record oriented, not stream.
> Files are created with a defined record format and record length.

Interesting. So the system actually "knows" where each field of a record 
is then?

>> What sort of access speeds do you get for reading or writing punched
>> card or tape?
> 
> As I recall about 60 characters per second.  The tape was paper, was 8 holes
> wide, and easily broken or scrunched.

Mmm, that's fairly fast for an optical system.

> On some of the older machines I worked with:
>   360/50 128KB
>   370/158 1MB
>   370/168 3MB
>   3033U 4MB
>   3081Q 8MB
> On the latest machine:
>   z/9 BC S03 - 16GB

I should probably go plot that on a graph against date or something... ;-)

>> Nice consistent naming scheme. Heh. :-/
> 
> You can always depend on IBM to be consistent.

Consistently inconsisten. ;-)

Still, some things never change:

   nVidia GeForce 8xxx
   nVidia GeForce 9xxx
   nVidia GeForce 2xx   <- WTH?

>> Now if I could figure out what a typical MIPS rating for a normal PC
>> today is, I'd have something dissimilar to compare it to. :-}
> 
> It is rather difficult to compare since they have such different architectures.

Yes, that's what I meant by "dissimilar comparison". It all rather 
depends on how many registers you have, how wide the various data busses 
are, what operations you can perform in hardware (e.g., is 
multiplication a hardware or software operation?), and so forth.

> I doubt that the internal processor speed is that much different than current
> PCs, but the i/o bandwidth is much higher in the mainframe.

Really? I didn't think anybody had mainframes any more... just big 
server farms.

> For example our two machines, which are on the small side for mainframes, are
> connected to a disk storage unit containing 15TB of data.  There are 4 high
> speed fiber optic channels from each processor to this unit. We can easily
> sustain i/o rates of over 10000 i/o operations per second, with peaks in the
> 20000 to 40000 range.

Sounds expensive.

> It is very rare for any single job to saturate the cpu.
> By running many jobs concurrently we make full use of the machine.

The *original* purpose of multitasking operating systems. ;-)

> If I could
> port POVRAY to our mainframe, I'm sure it would be slower than most current PCs.
> But cpu intensive work is not what the mainframe does best.

Out of curiosity, what *do* you use it for?

Post a reply to this message

> So the concept of a filesystem storing named files already existed at
> this time?

Yes, its been around long before I got into this business.

>
> Sounds expensive.
>

Its not cheap.  But for the volume of data we process, its not that much more
expensive than a server farm

>
> Out of curiosity, what *do* you use it for?

All company information is kept on mainframe disks and maintained by mainframe
programs.  All customer information, product information, personnel information,
sales information etc.  This company started in the early 1900s selling its
merchandise from a store.  Then expanding sales to mail order catalog, then via
telemarketing, and now by internet.  The web front-end is handled by a small
server farm, but all back-end work is done on our mainframe.

Isaac

Post a reply to this message

Invisible wrote:
> Does anybody know approximately when this time was?

1970's. By about 1980 most of those machines were likely retired.

> For that matter, does anybody have a broad timeline of when various 
> technologies were in use? What are the dates for things like core 
> memory, drum memory, punch cards, magnetic tape, relays, vacuum tubes, 
> transistors, ICs, etc?

Yes. Google.

> Was there ever a time when programs were entered into memory via 
> switches rather than some other medium?

Yes.
http://en.wikipedia.org/wiki/Front_panel

The first personal computers worked that way.
http://en.wikipedia.org/wiki/Altair_8800

> Was there ever a "punched tape" medium similar to punch cards?

Yes. That's why DEL is up at 127. Think about it.

> Similarly, you hear people talk about the VAX, the PDP, the varouis IBM 
> mainframes and Cray supercomputers. Does anybody know the timeline for 
> these, the technologies used and the basic design and performance details?

http://en.wikipedia.org/wiki/Computer_history

> (Sure, you can look up individual questions on Wikipedia, but the 
> articles tend to contain huge amounts of minute detail about specific 
> things. I'm trying to get a general overview of an entire era.)

So you want someone else to read the 10 pages and summarize it into 10 lines 
for you?  You need a job in management.

-- 
Darren New, San Diego CA, USA (PST)
    Ada - the programming language trying to avoid
    you literally shooting yourself in the foot.

Post a reply to this message

scott wrote:
>>> We added 1MB of memory to an IBM 370/168 in the mid 70s.  Price? $50000.
>>
>> Ouch! o_O
>>
>> It still somewhat blows my mind that you could do anything useful with 
>> so little memory.
> 
> You could do useful stuff with *far* less.

We ran an entire school district on 32K of RAM and 16M of online disk.
Core memory, so cycle time close to a milisecond, let alone nanosecond.

-- 
Darren New, San Diego CA, USA (PST)
    Ada - the programming language trying to avoid
    you literally shooting yourself in the foot.

Post a reply to this message

Invisible <voi### [at] devnull> wrote:
> Warp wrote:
> > Orchid XP v8 <voi### [at] devnull> wrote:
> >> Many hundred billion years ago, there existed a bizzare world that most 
> >> of us have only read about in books. Plastics did not yet exist, so 
> >> small children's toys were made of wood or metal, because there was 
> >> literally nothing cheaper.
> > 
> >   There were no children many hundred billion years ago.

> I was waiting for that one... *sigh*

  It's a so-called mathematician's answer.

-- 
                                                          - Warp

Post a reply to this message

Invisible wrote:
> So the concept of a filesystem storing named files already existed at 
> this time?

Generally, yes. But you usually wound up pre-allocating files, and they were 
contiguous on disk.

> Interesting. So the system actually "knows" where each field of a record 
> is then?

Records were fixed size, so it was trivial to calculate.

>> I doubt that the internal processor speed is that much different than 
>> current
>> PCs, but the i/o bandwidth is much higher in the mainframe.
> 
> Really? I didn't think anybody had mainframes any more... just big 
> server farms.

The people who want to do lots of I/O have machines where instead of GPUs 
they have IOPs.  A 800,000 line phone switch, for example, is pretty much 
all IOP, with something like a 68000 running the actual switching part.

Of course, what one might call a "PC" nowadays has a terabyte of RAM and 96 
quad-core processor chips, so the lines blur.

-- 
Darren New, San Diego CA, USA (PST)
    Ada - the programming language trying to avoid
    you literally shooting yourself in the foot.

Post a reply to this message

Invisible wrote:
> http://www.columbia.edu/acis/history/fisk.pdf
> 
> Hmm, interesting. Some of the guys I went to uni with learned COBOL. All 
> they could tell me about it is "if you miss out a dot, you get an error 
> message 300 lines later". Sounds delightful...

It's more fun to spell the first word wrong. "Idnetification Division." Then 
you get at least two error lines on each input line, or three if it's 
actually a top-level statement.

-- 
Darren New, San Diego CA, USA (PST)
    Ada - the programming language trying to avoid
    you literally shooting yourself in the foot.

Post a reply to this message

Invisible escreveu:
>> The first computer I ever saw was in 1997 at Glasgow University. It 
>> was an analogue machine that used valves.
> 
> 1997? o_O

just in time for the first version of GHC... :-)

-- 
a game sig: http://tinyurl.com/d3rxz9

Post a reply to this message

Invisible escreveu:
>> We added 1MB of memory to an IBM 370/168 in the mid 70s.  Price? $50000.
> 
> Ouch! o_O
> 
> It still somewhat blows my mind that you could do anything useful with 
> so little memory. Presumably for processing large datasets, most of the 
> data at any one time would be in secondary storage?

Large datasets then were also very tiny compared to large datasets of 
today. :)

>> When I was at university in the mid to late 60s, I took a FORTRAN 
>> programming
>> course.  Punch program into cards, place in card tray, come back 4 to 
>> 6 hours
>> later and pick up your card deck and output listing.  Fix your typos 
>> and bugs
>> and repeat the above process until the program works.

see the revolution that were programs like ed (and its successor vi) in 
bringing flexible terminal text editing rather than wasting tons of 
paper... :)

-- 
a game sig: http://tinyurl.com/d3rxz9

Post a reply to this message

Clarence1898 escreveu:
> All company information is kept on mainframe disks and maintained by mainframe
> programs.  All customer information, product information, personnel information,
> sales information etc.  This company started in the early 1900s selling its
> merchandise from a store.  Then expanding sales to mail order catalog, then via
> telemarketing, and now by internet.  The web front-end is handled by a small
> server farm, but all back-end work is done on our mainframe.

in other words:  mainframes were never supercomputers.  They were 
commerce-oriented machines rather than scientific number-crunchers.

-- 
a game sig: http://tinyurl.com/d3rxz9

Post a reply to this message

nemesis wrote:
> in other words:  mainframes were never supercomputers.  They were 
> commerce-oriented machines rather than scientific number-crunchers.

Some were.  Of course, "supercomputer" at the time had somewhat different 
connotations. But the mainframe I learned on had the optional "business 
unit" (i.e., support for COBOL instructions like BCD and block move) and the 
"scientific unit" (i.e., floating point hardware). You maxed that sucker out 
with four CPUs all with floating point hardware, and you could (and did) 
calculate orbits to the moon and such. It was a Sigma-9 Scientific Data 
Processor (aka Xerox-650 aka Honeywell <mumble>).

-- 
Darren New, San Diego CA, USA (PST)
    Ada - the programming language trying to avoid
    you literally shooting yourself in the foot.

Post a reply to this message

On 18/05/2010 6:30 PM, nemesis wrote:
> Invisible escreveu:
>>> The first computer I ever saw was in 1997 at Glasgow University. It
>>> was an analogue machine that used valves.
>>
>> 1997? o_O
>
> just in time for the first version of GHC... :-)
>

I thought that Andrew was the only one allowed to mention the "H" word. ;-)


-- 

Best Regards,
	Stephen

Post a reply to this message

On 18/05/2010 4:44 PM, Invisible wrote:
>> As I recall about 60 characters per second.  The tape was paper, was 8
>> holes
>> wide, and easily broken or scrunched.
>
> Mmm, that's fairly fast for an optical system.

Was it optical? I seem to remember it was mechanical with spring loaded 
teeth.


-- 

Best Regards,
	Stephen

Post a reply to this message

>> 1997? o_O
> 
> just in time for the first version of GHC... :-)

Actually about ten years *after* the first version of GHC.

Yes, I realise that sounds utterly absurd...

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

>> So the concept of a filesystem storing named files already existed at 
>> this time?
> 
> Generally, yes. But you usually wound up pre-allocating files, and they 
> were contiguous on disk.

OK.

Files only on disk? Or on tape too? (From what I've seen, punch cards 
didn't have this level of abstraction. It wouldn't be terribly necessary 
I guess...)

>> Interesting. So the system actually "knows" where each field of a 
>> record is then?
> 
> Records were fixed size, so it was trivial to calculate.

OK. But does the system know where the *fields* in a record are? Or just 
what size the records are?

>> Really? I didn't think anybody had mainframes any more... just big 
>> server farms.
> 
> The people who want to do lots of I/O have machines where instead of 
> GPUs they have IOPs.  A 800,000 line phone switch, for example, is 
> pretty much all IOP, with something like a 68000 running the actual 
> switching part.
> 
> Of course, what one might call a "PC" nowadays has a terabyte of RAM and 
> 96 quad-core processor chips, so the lines blur.

Yeah, I think the term "mainframe" is probably obsolete now. There are 
probably more exact ways to describe what type of computer you mean.

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

>> It still somewhat blows my mind that you could do anything useful with 
>> so little memory. Presumably for processing large datasets, most of 
>> the data at any one time would be in secondary storage?
> 
> Large datasets then were also very tiny compared to large datasets of 
> today. :)

Sure. But 1MB is such a tiny amount of memory, it could only hold a few 
thousand records (depending on their size). It would almost be faster to 
process them by hand then go to all the trouble of punching cards and 
feeding them through a computer. So it must have been possible to 
process larger datasets than that somehow.

> see the revolution that were programs like ed (and its successor vi) in 
> bringing flexible terminal text editing rather than wasting tons of 
> paper... :)

...not to mention card...

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

>>>   There were no children many hundred billion years ago.
> 
>> I was waiting for that one... *sigh*
> 
>   It's a so-called mathematician's answer.

Technically, 1 hundred billion years is about 7.3 times the estimated 
age of the universe, and would thus predate the Big Bang... if ou 
*really* wanted to split hairs. :-P

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

>> Does anybody know approximately when this time was?
> 
> 1970's. By about 1980 most of those machines were likely retired.

Right, OK.

>> For that matter, does anybody have 
> 
> Yes. Google.

Oh goodie.

>> Was there ever a time when programs were entered into memory via 
>> switches rather than some other medium?
> 
> Yes.
> http://en.wikipedia.org/wiki/Front_panel

Wow, OK. Looks like they have a dedicated switch for every individual 
memory register. o_O I was expecting a set of switches to punch in an 
address, and another set to input a datum. But I guess there are several 
designs possible... And hey, if you're spending £2,000,000 on a 
computer, what's a few thousand extra rocker switches?

> The first personal computers worked that way.
> http://en.wikipedia.org/wiki/Altair_8800

Mmm, interesting.

>> Was there ever a "punched tape" medium similar to punch cards?
> 
> Yes. That's why DEL is up at 127. Think about it.

Oh... oh dear god. You *are* kidding me, right??

>> Similarly, you hear people talk about the VAX, the PDP, the varouis 
>> IBM mainframes and Cray supercomputers. Does anybody know the timeline 
>> for these, the technologies used and the basic design and performance 
>> details?
> 
> http://en.wikipedia.org/wiki/Computer_history

Finally, a usable summary...

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Stephen <mca### [at] aolDOTcom> wrote:
> On 18/05/2010 4:44 PM, Invisible wrote:
> >> As I recall about 60 characters per second.  The tape was paper, was 8
> >> holes
> >> wide, and easily broken or scrunched.
> >
> > Mmm, that's fairly fast for an optical system.
>
> Was it optical? I seem to remember it was mechanical with spring loaded
> teeth.
>
>
> --
>
> Best Regards,
>  Stephen

I was thinking it was optical.  I don't ever remember having maintenance on the
read head.  But that's been over 35 years ago. Now I sometimes have a hard time
remembering what I had for breakfast.

Isaac

Post a reply to this message

>> Mmm, that's fairly fast for an optical system.
> 
> Was it optical? I seem to remember it was mechanical with spring loaded 
> teeth.

It wouldn't surprise me if both optical and mechanical systems have been 
devised at one time or another...

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Orchid XP v8 <voi### [at] devnull> wrote:
> >>>   There were no children many hundred billion years ago.
> > 
> >> I was waiting for that one... *sigh*
> > 
> >   It's a so-called mathematician's answer.

> Technically, 1 hundred billion years is about 7.3 times the estimated 
> age of the universe

  Depends on whether you are using American billions or European billions.

-- 
                                                          - Warp

Post a reply to this message

>>>>>   There were no children many hundred billion years ago.
>>>> I was waiting for that one... *sigh*
>>>   It's a so-called mathematician's answer.
> 
>> Technically, 1 hundred billion years is about 7.3 times the estimated 
>> age of the universe
> 
>   Depends on whether you are using American billions or European billions.

NAAAAAARGH!! >_<

I never did like Americans... (Then again, they probably don't like me 
either.)

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Orchid XP v8 escreveu:
>>> 1997? o_O
>>
>> just in time for the first version of GHC... :-)
> 
> Actually about ten years *after* the first version of GHC.
> 
> Yes, I realise that sounds utterly absurd...

Timing was just as innacurate as the original post.  All for the sake of 
a good joke. :)

-- 
a game sig: http://tinyurl.com/d3rxz9

Post a reply to this message

>>>> 1997? o_O
>>>
>>> just in time for the first version of GHC... :-)
>>
>> Actually about ten years *after* the first version of GHC.
>>
>> Yes, I realise that sounds utterly absurd...
> 
> Timing was just as innacurate as the original post.  All for the sake of 
> a good joke. :)

Hey, *I* had to look it up.

It still slightly frightens me that Haskell is actually this old... Just 
think how much better the world could be today if its ideas had caught 
on back then?

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Orchid XP v8 escreveu:
>>> It still somewhat blows my mind that you could do anything useful 
>>> with so little memory. Presumably for processing large datasets, most 
>>> of the data at any one time would be in secondary storage?
>>
>> Large datasets then were also very tiny compared to large datasets of 
>> today. :)
> 
> Sure. But 1MB is such a tiny amount of memory, it could only hold a few 
> thousand records (depending on their size). It would almost be faster to 
> process them by hand then go to all the trouble of punching cards and 
> feeding them through a computer.

that's not quite what Hollerith found with the American 1890's census. ;)

-- 
a game sig: http://tinyurl.com/d3rxz9

Post a reply to this message