🚀 go-pugleaf

In article <DLzvCC.JMD@artemis.sto.fdata.se>, Fredrik.Pettersson@sto1.wmdata.se writes:
>Dan Pop <danpop@mail.cern.ch> wrote:
>
>[snip]
>
>>Is it?  Pray tell, how is MacOS multitasking the following "application":
>
>>main() { for(;;); }
>
>I did one of those on a NT at work and it really shared processor time nicely,
>100% to that process and nothing to the other applications.
>
>So if you write your applications badly they will lock system more or less even
>in a preemtive environment.
>
>Is it my ignorance or could I be right on this point?

Probably ignorance.  

On a preemptive system, the endless loop above will get close to 
100% of the cpu *if it is the only thing ready to run*.  For example, 
on my Linux 1.2.13 machine, a single copy of the program gets >90% of 
the cpu:

  PID USER     PRI  NI SIZE  RES SHRD STAT %CPU %MEM  TIME COMMAND
17499 bobh      16   0    8  172  308 R    97.1  0.5  1:00 main


The number _varies_ though, depending on what other processes are
running.  Three copies of the program end up each getting 30%:

  PID USER     PRI  NI SIZE  RES SHRD STAT %CPU %MEM  TIME COMMAND
17499 bobh      30   0    8  172  312 R    32.5  0.5  4:05 main
17513 bobh      30   0    8  172  312 R    32.1  0.5  0:07 main
17514 bobh      30   0    8  172  312 R    32.1  0.5  0:06 main


At no time does the system "lock" or become the least bit non-
responsive.  I'm typing this as three copies of the "killer" 
program run.  Running 100 copies, OTOH, would certainly bog the
system down, as now everything would get 1/100 of the cpu instead
of 1/3.

So, yes, it is still bad form to write apps that busy-wait, even
on a preemptive system.  But one or two such apps won't kill the
system.

---
 Bob Hauck                                         bobh@wasatch.com
 Wasatch Communications Group                http://www.wasatch.com