[ale] a moment of silence please

[Jim Lynch]

Jim Philips wrote:
> The ever-expanding efficiency of data storage has got to be one of 
> the most dramatic technological changes in our lifetimes. I'm just 
> wondering if the processing power of our brains has caught up.
>
>   
My first computer (I didn't own it, NASA did) was a CDC 6600.  At that 
time it was the fastest computer in the world with a clock speed of 100 
ns.  It had a Bryant disk drive which was somewhat larger than a large 
freezer and had a whopping 75 M bytes (CDC bytes were 6 bits wide). 

It had a very large memory, 131,072 60 bit words.  Core memory no less.  
What was unique was that the CPU had no I/O capabilities at all.  I/O 
was performed by 10 (sorta) peripheral processing units (ppu) that were 
attached to things like printers, disks, card readers, card punches, 
tape drives, etc.  Actually there was only one ppu but it was in a 
"slot" so it accessed I/O and ppu memory for a slice of time. 

The operating system at that time didn't have a memory scheduler.  You 
loaded "jobs", or decks of card into the system and the program got 
loaded into memory.  The operator could swap the job out to disk 
manually and load another program, and another, etc.  However the 
scheduler was the operator, who periodically swapped out/in various 
programs depending on priority of the user/job. 

Fortran and assembly were the only languages in use, generally.  There 
was a fortran compiler named "run" that was popular. 

Later we upgraded the disks to 808's.  These were boxes that were 10 
feet long, 7+ feet high and 3 feet thick.  Each half of the cabinet held 
a set of platters with a large motor to spin the platters.  The shaft 
was vertical with the motor in the middle with about 8 platters below 
and 8 above the motor.  The heads were on a positioner located inbetween 
the motors and were hydraulically driven.  The mechanism was symetric, 
so that when the heads went out on one side they also went out on the 
other the same amount.  We had to maintain the disk surfaces by taking 
what amounted to large cue tips (the kind you use on your ear wax), 
alcohol and water and scrubbing them while they were turning.  Head 
crashes were  pretty exciting. 

I'm a little hazy on this, but I think the first OS didn't allow for 
permanent file space on disk.  If you wanted to save your data, you had 
to write it to tape.  The disk space was for temporary data only.  These 
were real memory systems meaning the program had to reside completely in 
memory while running.  The only swapping was done by the operator and 
that was to completely move a program one way or another.

There was CPU scheduling.  If there were two or more programs that fit 
into memory, then they both would be candidates.

So I didn't get to fiddle with switches, except to boot the system, you 
had to put the boot code into the ppu via an array of 15x6 switches.   
That would tell the system which device to boot from.  The options were 
7 track magnetic tape or the card reader.

Jim.