silverdr_at_inet.com.pl
Date: 2006-03-02 10:23:21
On 2006-03-02, at 09:44, Spiro Trikaliotis wrote:
> Hello,
>
> * On Wed, Mar 01, 2006 at 01:33:10PM +0100 silverdr@inet.com.pl wrote:
>
>>> massive power of nowadays processors. I wonder if it could be done
>>> with, let's say, a 100 MHz Pentium-1.
>>
>> It can be done even on an Amiga 500. The trick is to use well timed
>> and predictable interrupt handling rather than bit polling.
>
> Which modern, non-RT OS gives you preditable interrupt handling?
> Neither
> Linux nor Windows do. This is the first big challenge for
> cbm4win/cbm4linux.
You're right. There is a challenge - not even one (that's why most
solutions are still based on stone-age OSes). But the point was that
it can actually be done in a multitasking environment on machines of
far less CPU power than some people believe. I remember for example
some early problems with serial communication on the Amigas, whenever
the communication wouldn't be reliable due to timing problems once
the more advanced schedulers with dynamic priorities were introduced.
Things have to be handled appropriately even up to the level of forbid
() and more but I don't agree with the statement that required timing
(in a multitasking env) can be achieved only with massive computing
power of todays CPUs and answer the question "if it could be done
with [...] a 100MHz Pentium-1". The answer remains "yes", even if
there are conditions to that. ;-)
--
[...] the truth has finally come out: Apple Computers promote Godless
Darwinism and Communism - Dr. Richard Paley
Message was sent through the cbm-hackers mailing list
Archive generated by hypermail pre-2.1.8.