On Monday, November 17, 2003, at 10:04 AM, Brian Chase wrote:
>>    Well I don't remember where I got 25MHz, it's been in my head for 
>> at
>> least a decade.  But I can tell you for certain that the 78032 on a
>> KA630 is clocked at 20MHz, as I've overclocked a couple of them...so
>> that would seem to support the idea, as I doubt its successor would
>> have a lower clock rate.
>
> Maybe the KA630s had a crystal oscillating at 20MHz, but the CPU cycle
> time is listed as being 200ns (5MHz) in the VARM.  DEC most probably
> divided up that 20MHz crystal output, using only a fraction of it for
> CPU timing.  It'd certainly be interesting to know what was going on at
> a very detailed hardware level, but I'm not much of a computer 
> engineer.

   Do you mean the machine cycle?  Some processors (mostly CISC ones) 
take multiple clock cycles to execute an instruction, and some take 
varying numbers of clock cycles depending on the instruction.  As an 
example, the Intel 8051 (I mention this one because I happen to be 
doing some cycle counting on an 8051-based system at the moment) uses 
12 clock cycles per machine cycle, and anywhere from one to three 
machine cycles per instruction.

   The oscillator on the KA630 is a 40MHz part.  As I seem to recall 
from a thread here a few years ago, it is divided by two for the rest 
of the board, and then split into an eight-phase clock within the 78032 
for use internally.

        -Dave

--
Dave McGuire                      "My tummy hurts now, but my soul
St. Petersburg, FL                 feels a little better."     -Ed