To: [EMAIL PROTECTED]
>From: Alastair Lee Sumner <[EMAIL PROTECTED]>
>Subject: [rtl] Scheduler implementation
>Date: Sun, 10 Oct 1999 15:00:53 +0100 (BST)
>Reply-To: [EMAIL PROTECTED]
>
>Hi
>
>Would it be true to say that all the data structures and information
>needed to implement a simple scheduling policy eg. rate monotonic,
>possibly deadline monotonic, are contained within the default scheduler
>source, so that by modifying the default scheduler (beta16) you could
>implement another policy? No SMP or FPU to keep it simple for the time
>being.
Rate monotonic I have heard of. Deadline monotonic is something new to
me. Perhaps Alstair should explain what these terms mean and not
assuming that everybody have learned computer "science" jargon,
particularly hardware designers might be reading rtl email. I am sure
many hardware designers also must do the programming without the benefit
of software experts.
There are many "hard" real-time control systems that may have one task to
train a sensor/seeker tracking a moving object while another task
controlling some motion of the platform with its own position and rate
sensors. Generally there the amount of information communicated between
these two tasks are not large. Could each task be programmed separately
and determine the actual time each task takes? This require no fancy
algorithms which cannot help to create time anyway. The time taken by
rt-os need not be a concern if the sum exceeds real time available
already. Also, did the teachers suggest the use of individual processor
for each task as alternative and compare their system cost, power, size
and the amount of rtlinux facilities they need as a designer
(hardware/software) should?
>I'm investigating the feasibility of learning the rt/kernel level code in
>a short period of time, for an undergraduate final year project on hard
>real-time scheduling.
I am curious to know if Alstair's teachers are experienced designers of
"hard real-time" systems or have been brought up under time-shared
mainframes with multi-tasking operating systems. The use of the word
scheduling implied using a computer in a time-shared manner. Did they
explain why a control system need to assume a time shared computer?
These are merely questions to satisfy my ignorance and are no reflection
on Alex or his teachers.
SY Wong 10.10.99
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
