[OS X TeX] (OT) Tiger and not Tigrrrr :-)?

[Michael McCracken]

On Jul 27, 2005, at 11:22 AM, Peter Dyballa wrote:

>
> Am 27.07.2005 um 16:46 schrieb Michael S. Hanson:
>
>
>>     Just an FYI, which you may wish to take with a grain of salt:   
>> In a recent meeting to discuss ways to deploy and use Xgrid at my  
>> campus, a local tech support guy noted -- and the Apple rep on  
>> hand confirmed -- that the Mach kernel doesn't actually respect  
>> 'nice' settings.  The tech support guy called it a "placebo":  the  
>> user sees the nice setting (say in 'top'), but the kernel  
>> (reportedly) completely ignores it.
>>
>
> In other UNICes nice too isn't what it was ten or 15 years ago.  
> I've found that a nice'd process cannot make the CPU fan start on a  
> portable -- i.e. the processor seems not to be switched from  
> minimal to maximal speed.
>
> For me and my PowerBook this is a valuable feature: I can safely  
> start long compilations before I go to sleep (make bootstrap for  
> Emacs, fink update-all) or visit the public swimming pool to cool  
> down myself ... and when I return revived I have a new FontForge or  
> such!

I've also seen measurable differences when changing task priority  
using 'nice'.
A possible explanation for the tech support guy's assertion is that  
the darwin task/thread priority mechanism is complicated enough that  
it doesn't map well to what you learn about unix in school.
(At least, not what *I* learned about unix in school!)

According to the docs, there are separate scheduling 'bands' as well  
as priorities within that band.
This is why programs can hog the CPU but easily not make the mouse  
movement choppy, or skip iTunes, for instance. When you use nice, it  
puts that process into a separate band. The docs don't say  
explicitly, but it's implied that the order of band priorities goes:  
graphics server, dock, background apps, foreground apps, (re)nice'd  
processes, then other.

OK, that's enough offtopic OS stuff from me. If you want to see where  
I got this from:
http://developer.apple.com/documentation/Darwin/Conceptual/ 
KernelProgramming/scheduler/chapter_8_section_4.html

FWIW, I've seen some bad performance from spotlight too.

-mike

--
     Michael McCracken
     michael_mccracken at mac.com
     http://michael-mccracken.net/


--------------------- Info ---------------------
Mac-TeX Website: http://www.esm.psu.edu/mac-tex/
           & FAQ: http://latex.yauh.de/faq/
TeX FAQ: http://www.tex.ac.uk/faq
List Post: <mailto:MacOSX-TeX at email.esm.psu.edu>