> > The problem is context switch time. With Hard real time systems
> > it is <10 microseconds. I am trying to find out context switch times in
> > Linux also. There will be several processors and DSP's
> > many pegs many holes
> First of all , in my guess, 10 microseconds is really hard!.
> One of the solution for this HARD realtime application could be a
> bufferring scheme. This idea is simple, add another single chip
> (like scenex? 100 mips PIC) doing simple processing
> and do main processing at 1 millsecond interval. But this is
> only useful for periodic request.
On a normal Pentium-class computer, RTLinux folks measure latencies
of about 2 microseconds typical. There is a tail, though. Look
at the Linux Journal article about flying a radar set into a hurricane.
The final extreme is to throw a gate array at the problem. That
can be very inexpensive and general, just imagine Linux running
on the same gate array, see http://www.opencores.org/ OR1K project.
That would give you the flexibility to put the "right" amount of
reflexes in hardware (often that means 10 +/- 0.1 ns delay), and
the remainder in c code, with arbitrary coupling between the two.
- Larry Doolittle <LRDoolittle@lbl.gov>
This message resent by the email@example.com list server http://www.uClinux.com/
This archive was generated by hypermail 2b30 : Sun Apr 07 2002 - 00:01:34 EST