T
Tom Gardner
Guest
On 14/08/20 08:27, Tauno Voipio wrote:
Ach, the 276 was a typo on my part; I remember it being
a 2kIPS machine, i.e. 576us. I don\'t think I was aware
of the 288us timing.
ISTR I used the instruction modifier bit in my hand-assembled
program. That converted from one 5 channel paper tape code
we had at my school to the Elliott 5 channel code. (ASCII was
wonderful
)
I forget the details, but it had two \"states\" (for fig/num shift)
and a computed goto based on the next character read. Someone else
tried to do it with if-the-elses, and failed miserably.
Although I didn\'t realise it at the time, that taught me that
/thinking/ and working out the right abstraction makes things
much more tractable.
Youngsters don\'t seem to realise that, and just use whatever
they\'ve been taught.
On 14.8.20 9.00, Tom Gardner wrote:
On 14/08/20 00:54, ke...@kjwdesigns.com wrote:
On Thursday, 13 August 2020 at 13:59:08 UTC-7, Tom Gardner wrote:
...
The 1802 wasn\'t too odd, except that the PC was designated
by a 4-bit register under programmer control.
I agree on Hoare\'s Algol. It was a huge step forward from Autocode
and direct octal code input.
I did both, trimphantly reinventing a simple FSM in the latter case
The Algol compiler fitted in 4Kwords. I met Hoare once, and he
was surprised when I mentioned having used it
The first computer I used was an Elliott 803, at Bangor University, with 8k
words (it had a second cabinet for the additional 4k). We ran Algol with it
which was loaded from paper tape, a 6\" diameter roll. The run time library
was on a second slightly smaller roll of tape and had to be loaded with the
executable from the compiler.
That machine had a 39-bit word, and used a serial architecture running at
~145kHz clock rate, with OC45 transistors giving about 0.003 Mips IIRC.
I used one while was in the 6th form, in the neighbouring Ewell
Tech (now NESCOT).
I\'m not sure whether they loaded Algol from the paper tape, since
it also had the (sprocketed) magnetic film devices.
The instruction cycle time was 276us.
I almost went to Bangor, but Southampton was more convenient.
If you are ever near TNMoC, go and see one working, and listen
to it playing music (the high notes are very flat!). When I
mentioned I was an electronic engineers and had used one,
they whipped out the schematics and we discussed them.
Now that\'s what I call a /good/ museum. By comparison, Bletchley
Park next door is a bog-standard museum only worth seeing once.
The basic cycle was 288 us, and most instructions used two cycles,
576 us. IIRC, the only single-cycle instructions were control transfers.
There were two instructions in a 39 bit word, 6 instruction code bits
and 13 address bits for 19 bits per instruction. The extra bit in
the instruction word was an address modifier bit: If it was on, the
address of the second part was indexed with the result of the first
part before use.
Ach, the 276 was a typo on my part; I remember it being
a 2kIPS machine, i.e. 576us. I don\'t think I was aware
of the 288us timing.
ISTR I used the instruction modifier bit in my hand-assembled
program. That converted from one 5 channel paper tape code
we had at my school to the Elliott 5 channel code. (ASCII was
wonderful
)I forget the details, but it had two \"states\" (for fig/num shift)
and a computed goto based on the next character read. Someone else
tried to do it with if-the-elses, and failed miserably.
Although I didn\'t realise it at the time, that taught me that
/thinking/ and working out the right abstraction makes things
much more tractable.
Youngsters don\'t seem to realise that, and just use whatever
they\'ve been taught.