On Sun, Jan 02, 2000 at 03:13:10PM -0800, David O'Brien wrote:
> On Sun, Jan 02, 2000 at 04:53:55PM -0600, Karl Denninger wrote:
> > On Sun, Jan 02, 2000 at 11:49:08PM +0100, Poul-Henning Kamp wrote:
> ..snip..
> > > If you intend to keep up this "sour grapes" attitude, despite all
> > > the helpful answers you have gotten so far, you should consider
> > > stopping before you have worn out your welcome.
> >
> > Go fuck yourself Poul.
> >
> > Yes, the "F-word".
> > You have no ability to determine "my welcome".
>
> Karl you are one fscking asshole. As a member of this list, you *HAVE*
> worn out your welcome in my inbox.
So what?
> You asked Poul-Henning about *good* timekeeping HW. He told you. $600
> isn't that much for computer server related hardware. Warner seconded
> the recommendation.
So what?
> You pay less, you don't get the best. Its up to you to weigh your needs
> vs. what you are willing to pay.
Shilling for a difference that is irrelavent is horseshit. It amounts to
lying. Statistically these two time sources are almost EXACTLY equivalent.
> And you want to bitch?? What is your fscking problem?
Its a bullshit argument David and is nothing other than BLATENT shilling
for Motorola - at someone else's expense!
Without a specialized board that counts these nanosecond ticks from the
time it gets the PPS signal to the time the latch is read this kind of
accuracy claim is worth exactly NOTHING.
WITHOUT it, which is EXACTLY what REAL WORLD use of these devices is
going to be running under, the impact is ZERO.
Further, the ACTUAL impact in the real world of time stability to the
dozen-nanosecond-range is ALSO zero, with the possible exception of
physical control processes in the nuclear research field (such
applications are NOT running on "standard" Unix machines!)
Finally, in the world of sync'ing time between systems (say, in a multiple
server cluster) that kind of precision is ALSO worth ZERO!
NOBODY with off-the-shelf hardware can obtain repeatable 10ns results
from a standard Unix machine.
NOBODY.
And THAT is a fact, easily determined through nothing more than Intel's
databooks on their processors and the interrupt response time they exhibit.
--
--
Karl Denninger ([EMAIL PROTECTED]) Web: http://childrens-justice.org
Isn't it time we started putting KIDS first? See the above URL for
a plan to do exactly that!
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
