lkml.org 
[lkml]   [1997]   [Sep]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [patch] QNX-style scheduling v1.10 for Linux 2.0
Hi!

> That not true, compiling the kernel doesn't use much ram.. (Make -j2 shows
> a loss of about 7megs free before adding buffers).. And with the QNX
> patch, a reallyniced task just plain doesn't run if there are other tasks
> that want to run.. So they shouldn't induce swapping if there are other
> tasks doing anything (because the compile will not get ANY cpu if there
> are other tasks)..
>
> And of course, you can still nice tasks..
>
> FYI- I've used the QNX patch on a 486-dx33 with 8megs.. And I can compile
> the kernel in the background with almost no noticeable performance hit..
> (Without the patch the system is awefully slow).. Please try patches, or
> atleast understand how they work before you critisize them.. If you dont
> you will likely be incorrect.. I've tried it, it works. The only

FYI, I'm running QNX v1.06 on 2.1.55 with 16Meg of ram. ANd since I'm
not that kind of person that eats 100% of cpu time, it does generate
some slowdown. And yes, it's a bit better with QNX. And no, QNX can
not help it because tasks are waking up after next timer tick - 10msec
too late.

> performance hit is from the disk activity.. But if I'm trying to use the
> computer, the compile doesn't get much cpu, so it doesn't geneate as much
> disk activity..

Pavel

--
--
This is my little buggy signature... Pavel
GCM d? s-: !g p?:+ au- a--@ w+ v- C++@ UL+++ L++ N++ E++ W--- M- Y- R+

\
 
 \ /
  Last update: 2005-03-22 13:40    [W:0.035 / U:0.088 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site