[Info-vax] AlphaServer DS10 470 MHz?

[VAXman- at SendSpamHere.ORG]

In article <jfh1s0$1c3$1 at Iltempo.Update.UU.SE>, Johnny Billquist <bqt at softjar.se> writes:
>On 2012-01-22 13.13, Johnny Billquist wrote:
>> On 2012-01-22 05.37, Michael Moroney wrote:
>>> MG<marcogbNO at SPAMxs4all.nl> writes:
>>>
>>>> An unusual occurrence, for me at least, as I was presented with the
>>>> following, below:
>>>
>>> [470 MHz vs. 463 MHz]
>>>
>>> Digital exaggerated the speed ratings of some of its Alpha products
>>> slightly for some (marketing, I assume) reason.
>>
>> The measured cycle time is 2.16 ns. How accurate do you think you can
>> measure cycle time? What is the cycle time to get to 470 MHz?
>> Just do the math, and realize that the measurement of cycle time can't
>> be done with enough precision to get the absolute right number.
>
>FYI: 470 MHz gives a cycle time of 2.12766 ns. 463 MHz gives a cycle 
>time of 2.15983. That is a difference of 32.17 ps. 32 picoseconds is 
>pretty little. Is the accuracy of the measurements enough to deal with 
>such small times?

I believe that, as someone else has pointed out, it's just for marketing.
It's a small round up to call it a DS10 470MHz instead of a DS10 463MHz.


>All that said, I do not know if the stated 470 MHz is actually true. 
>Anyone checked the actual crystal or clock generator circuitry to see 
>what is actually the frequency base?

I suspect that it's far easier to accurately measure the clock speed than
it is to try to measure the zero-crossings accurately for cycle-time.  As
you know, they're just reciprocals of each other.  How far do you want to
carry out the decimal???

-- 
VAXman- A Bored Certified VMS Kernel Mode Hacker    VAXman(at)TMESIS(dot)ORG

Well I speak to machines with the voice of humanity.