First computer with real-time clock?

From: Paul Koning <pkoning_at_equallogic.com>
Date: Mon Aug 2 08:19:52 2004

>>>>> "Vassilis" == Vassilis Prevelakis <vp_at_cs.drexel.edu> writes:

 Vassilis> Sellam Ismail (vcf_at_siconic.com) wrote:
>> Does anyone know what was the first computer to have a built-in
>> real-time clock?

 Vassilis> I take it you mean non-volatile clock like the M48T02 RTC
 Vassilis> chips on the early SUN-4 machines.

 Vassilis> The reason why there is a difference, is that most
 Vassilis> operating systems (esp. those that have preemptive
 Vassilis> schedulers) need a regular timer interrupt. So they use
 Vassilis> that interrupt for their time-of-day clock. This method is
 Vassilis> generally accurate and convenient since mainframes were not
 Vassilis> likely to be switched off overnight.

 Vassilis> Another advantage held by the early mainframes is that they
 Vassilis> were kept in a stable environment, which meant that their
 Vassilis> clock keeping was pretty good anyway.

That depends.

If the interrupts are derived from the mains power, then the
timekeeping is usually extremely good -- except if the mains is a
local generator not connected to the grid, in which case it is likely
to be very poor.

Conversely, if it's derived from a crystal oscillator, the timekeeping
often was quite bad. The reason is that ordinary crystal oscillators
(except for those used in watches) come with an 0.01% tolerance, which
means an error of over a minute per day. Since actual error
distributions on crystals are bimodal, not Gaussian, you would
normally get just about that error.

This is why DEC machines with a KW11-P were typically run at line
frequency and not off the 100 kHz oscillator.

          paul
Received on Mon Aug 02 2004 - 08:19:52 BST