First computer with real-time clock?

From: Hans B PUFAL <hansp_at_citem.org>
Date: Mon Aug 2 01:11:45 2004

Vintage Computer Festival wrote:

> As far as need to keep the time during power off, I don't know that in
> this case it would matter too much. It would just have to be a
> combination of hardware and software that kept the current date in the IBM
> 650. The 650 was small enough that you could power it down during off
> hours, though I don't know what actual practice was (perhaps it was too
> much of a chore to boot up so it was kept on all the time). At any rate,
> as long as the 650 had some way to generate uniform time pulses, a
> real-time clock/calendar could be implemented, but this would probably
> have to be at a hardware level since the 650 was primarily a batch
> machine, or it would have to be programmed to store the date/time in some
> memory location that the other applications would not touch, and then some
> counter would have to be implemented that could update a memory location
> with the number of ticks since the last access or whatever.
>
> Of course, as I've been suggesting, there could have been a totally
> separate clock/calendar device (heck it could have even been mechanical)
> that kept the local time/date and then transferred it to the 650 upon
> request. Just like the clock in a modern PC.

OK, I've sent this question to the IBM Haus of History in Sindelfingen
in Germany. They have a working 650 so should be able to provide an
answer. I'll keep you informed.

The periodic interrupt (50 or 60 times a second depending on which side
of the pond you were) was most definitely available in 1964 on the IBM
System/360. At each tick, the hardware decremented a word in memory and
when it overflowed to zero, a program interrupt was invoked. The handler
would do whatever it needed and reset the memory word.

Earlier IBM machines also some form of this feature. On the 7094
modified for CTSS it was an IBM special feature.

Regards,

   -- HansP
Received on Mon Aug 02 2004 - 01:11:45 BST