Re: uptime not reporting CPU usage on Windows 7 (Possibly only when running in VMWare)

[Andrew DeFaria <Andrew at DeFaria dot com>]

 On 12/31/2010 01:11 AM, Andy Koppe wrote:
> > IMHO it's 100% better than just outputting 0's. Putting out 0's gives you no
> > info at all!
> Bollocks. You'd be the first to complain that those stupid Cygwin devs
> don't even understand what an average is.
Hold on there Tonto! I said nothing of the sort. If you're gonna put 
words in my mouth then please refrain for claiming that I insulted 
somebody else in the manner you did above. I didn't say that the Cygwin 
devs didn't understand what a loadavg is and I hardly called them 
stupid. You should be ashamed of yourself for making such a leap of 
misjudgment and then pinning it on me!
> The 0% tells you pretty
> clearly that that information is not available.
I beg to differ. 0% is indistinguishable from "the machine was not busy 
at all". IOW it *could* say "that information was not available" and it 
*could* as easily say "the load avg was actually 0%". You can't tell so 
it's hardly "pretty clearly". Indeed reporting say a negative number 
would impart more information as the user would say "Huh?!?" and then 
perhaps the man page could explain that a negative number is returned 
because this information is not easily obtainable by Cygwin under 
Window. But reporting 0 is by definition ambiguous!
> > I beg to differ. I don't see how having 0 values is better than some
> > approximation of load. Surely Windows has some measurement of the number of
> > processes in the run queue.
> And surely you'd be able to find it at http://msdn.microsoft.com/library.
Thanks for pointing out where they can find the info.
> > A simple count of the number of processes with CPU usage>  0
> > (minus the system idle) process would be a good start.
> Hardly.
This is surely a subjective opinion. You're opinion is "hardly". Mine is 
"good start".
>   The challenge here is to actually collect the data for the
> last 15 minutes, so you'd be looking either at some sort of service
> process waking up frequently to sample the CPU state, or some unholy
> and unreliable scheme where Cygwin functions go off and do that every
> now and then before doing their actual job.
While any of the 1, 5 or 15 minutes would be challenging as you say 
here, I still maintain that an instantaneous measure to substitute for 
at least the 1 minute figure would be an improvement over just putting 
out 0. You are free to maintain a different opinion just as I am free to 
hold this one.
--
Andrew DeFaria <http://defaria.com>
Give a man a fish and he'll eat for a day, teach a man to phish and 
he'll suck your bank account dry


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple