Nicer than nice! (Retraction)

[Uday Hegde]

In article <90Feb5.165122est.5069 at neat.cs.toronto.edu> lamy at cs.utoronto.ca (Jean-Francois Lamy) writes:
>someone wrote:
>>> Is there a possibility to change the priority of a process more than a

>and Uday Hegde confuses matters by saying:
>>process with that nice value [-19] runs only when nothing else in the system 
>>wants to.
>>Correct me if I am wrong.
>
>Indeed.  The scheduler uses niceness as one of many factors in determining
>priority.  In *practice*, several maximally niced jobs can still impact

Sorry! I was quoting the man pages for "nice" and "renice(8)" on ULTRIX when I
posted on the net. The man pages clearly states that nice value 19 is 
different from the rest and is useful when you want to ensure that the process
runs when nothing else in the system wants to. However, having actually 
experimented with different nice values, it appears that the man pages are 
wrong. My apologies for "causing confusion". :)

______________________________________________________________________________
Uday Hegde.                          internet: hegde at cs.umass.edu         
UMass, Amherst.                                hegde at umvlsi.ecs.umass.edu
Phone: (413) 545-3899                 bitnet:  hegde at umass.bitnet    
_______________________________________________________________________________