Tech Support > Computer Hardware > Microprocessors > real time embedded system requirement on processor architecture
real time embedded system requirement on processor architecture
Posted by joggingsong@gmail.com on March 7th, 2008


Hi,
For a real time embedded system, is there any requirement on the
processor architecture to meet the real time
requirement? Does realtime system depend on realtime OS to meet the
realtime requirement?

Thanks
Jogging

Posted by Vladimir Vassilevsky on March 7th, 2008




joggingsong@gmail.com wrote:

Yes. The CPU must be in stock at Digikey, otherwise the realtime could
be problematic.

Looks like your professor gives you a hard realtime.


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

Posted by Jack Klein on March 7th, 2008


On Thu, 6 Mar 2008 17:35:43 -0800 (PST), joggingsong@gmail.com wrote
in comp.arch.embedded:

What is "the real time requirement"? I did not know that there was
only one. Every real time system I have worked on so far, and there
have been a lot of them, have had their own different requirements.

Now that we've straightened that out, yes, just about every real time
system usually impose some minimum performance requirements on the
processor/controller/DSP.

The requirements of some real time systems can be met by an original
8051 running at 12 MHz. The requirements require much more powerful
processors.

Some do, some don't.

--
Jack Klein
Home: http://JK-Technology.Com
FAQs for
comp.lang.c http://c-faq.com/
comp.lang.c++ http://www.parashift.com/c++-faq-lite/
alt.comp.lang.learn.c-c++
http://www.club.cc.cmu.edu/~ajo/docs/FAQ-acllc.html

Posted by Tim Wescott on March 7th, 2008


On Thu, 06 Mar 2008 17:35:43 -0800, joggingsong wrote:

"Real time" is pretty much accepted to mean that if an event happens, and
your system doesn't respond to it quick enough, then the system isn't
meeting spec. Some folks separate this into "hard real time" which means
that if even one deadline is missed by even one nanosecond then disaster
strikes, and "soft real time" which means that some deadlines can be
missed, with some sort of graceful failure that gives degraded
performance long before things crash and burn.

So the only requirement on a processor architecture to be suitable for
_some_ real time application is that it be deterministic, that is that
you can always calculate the maximum possible time for the processor to
carry out a given task. Once you've established that, then you can
determine if it'll be fast enough to meet _your_ particular real-time
requirement.

make the development easier. (note that a real time OS also makes the
development more complex and the problems more obscure -- developers of
really critical applications tend to shy away from RTOS's for this
reason).

Regardless of what framework you're doing it in, real-time software
development requires that you pay attention to details from one end to
the other. It's not the right place to jump in and play without knowing
what you're doing.

--
Tim Wescott
Control systems and communications consulting
http://www.wescottdesign.com

Need to learn how to apply control theory in your embedded system?
"Applied Control Theory for Embedded Systems" by Tim Wescott
Elsevier/Newnes, http://www.wescottdesign.com/actfes/actfes.html


Similar Posts