Re: Real Operating Systems That Take Control

From: f0dder (f0dder_spicedham_at_flork.dk)
Date: 07/08/04


Date: Thu, 8 Jul 2004 14:34:55 +0200


> So "real operating systems" grab complete control of all processes
> and limit the time the processor can spend with each process ? I
> can see something like that being effective, but I can also see that
> it could slow down disk operations and screen writes and possibly
> other hardware i/o.
>
On NT, you have to have correct privileges to change your process and/or
thread priorities to high and realtime. If you don't have this privilege,
the call won't fail, but you'll only get the highest of the "dynamic"
priorities - which is not enough to lock up the system.

When a thread runs at realtime and eats up the CPU, other realtime threads
won't be scheduled - but IRQs are still processed. If you're interested
in learning more about this, have a look at Solomon and Russinovich's
"inside windows 2000", it's a very good read if you want to design well
working applications for windows. And an interesting read even if you're
just a "power-user" :-)

> I'm very much interested in this topic, and I'm thinking that there
> would be have to be quite a bit of memory buffering to speed
> hardware i/o up.
>
Unfortunately, windows is a bit conservative wrt. the use of filesystem
cache :(. There's a registry setting (LargeSystemCache=1, on XP and more
recent you can even set this from the GUI), which works great, but
unfortunately I can't use it after I got a radeon card - the ATi catalyst
drivers have some _serious_ issues with this setting that can end up in
_massive_ data corruption (and almost cost me 30+ gigs of data.)

> I'm curious about what a real operating system is... are you
> referring to Linux or Unix ?
>
NT, BSD, Mac OS X, (to some degree) linux, springs to mind. Win9x certainly
ISN'T a real OS, it's a monstrously overgrown dos extender full of 16bit
code and no security... but it does the trick on older hardware :)
(Of course NT4 runs perfectly on a mmx200 w/64 megs of ram, and that does
qualify as 'old' to me, and I'd certainly rather be on NT4 than win9x...
but to each his own)



Relevant Pages

  • Re: One SBC or two?
    ... We're looking at a major upgrade of an existing ... x86 SBC to do a lot of realtime and ... realtime stuff, ... connectivity and Windows is a bit problematic. ...
    (comp.arch.embedded)
  • Re: What could be the best for Timing
    ... Depends on what they mean by "timing", which is part of why I asked about wall-clock time. ... Windows is not a real-time system, and it is an illusion to pretend that it is. ... time" or "meeting hard realtime deadlines". ... If good precision 99.9% of the time is good enough, ...
    (microsoft.public.vc.mfc)
  • Re: C++ Question about its possibilites to XP..
    ... Regardless of programming language. ... impossible to write realtime applications for generic off-the-shelf PCs ... applications with nifty windows and stuff. ... as it can get with windows - soley depending on your skill and training. ...
    (comp.lang.cpp)
  • Re: E-TRUST ANTIVIRUS
    ... > We are having this problem on a windows 2000 terminal server build. ... The problem I am having is that as long as Realtime ... > monitoring is disabled, the server stays up but as soon as realtime ... I've been running eTrust EZ Anti-Virus on a Windows ME and two Windows 2000 ...
    (microsoft.public.win2000.general)