[OT]: Quiz for scheduler "want to be authors" [Was]: Re: code of process scheduling
Andrey Andreev
andreev at cs.helsinki.fi
Fri Oct 8 11:56:03 UTC 2004
Nifty Hat Mitch wrote:
> On Thu, Oct 07, 2004 at 01:10:44PM +0100, James Wilkinson wrote:
>
>>Nifty Hat Mitch wrote:
>>
>>>Quiz for scheduler "want to be authors":
>>>
>>>Given:
>>> A two man racing team.
>>> Race is 100 Km in length.
>>> The first team member drives the first half at 50Km/hr.
>>>
>>>Question 1:
>>> How fast must the second driver drive the second
>>> half of the race for the TEAM to average 100Km/hour?
>>
>>Andrey Andreev (correctly) answered:
>>
>>>The only way the team would average 100 km/h along the length of the
>>>race is, as far as my imagination goes, the unlikely event that the
>>>second driver quantum teleports to the finish line instantly. Either
>>>that, or fall into a conveniently placed wormhole or something.
>>>
>>
>>Mitch's second question was:
>>
>>> Essay, How the heck does this apply to the Linux scheduler.
>>
>>Andro asked:
>>
>>>Seriously, how the heck does this apply to the Linux scheduler???
>>
>>One possible drawback with a scheduler is that it spends so much time
>>trying to work out the Right Thing To Do that it never gives the system
>>the chance to do it: gettting on with something vaguely productive might
>>get things done faster.
Good point.
>
> Hint... look at the thread for
>
> k3b Burn Speed
Silly me, I can't see the connection. Can I bother you to elaborate? Or
are we getting too far with the offtopic discussion?
>
> And yes computation for scheduling can get out of hand.
> Simple and "fair" what ever that is works...
Regards,
//Andro
--
Andrey Andreev
University of Helsinki
Dept. of Computer Science
More information about the fedora-list
mailing list