Interrupt latency problems on MPC860

[Till Straumann]

Phil Torre wrote:

> We are using RTEMS 4.5.0 on a 50MHz MPC860.  Our application relies
> on a particular periodic interrupt being serviced with no more
> than 78 microseconds of latency.  We keep having problems because
> that interrupt isn't being serviced on time.
>
> There's not much else going on in the system currently.  Several
> tasks are asleep or blocking on I/O, but none of them should be
> waking up and causing a context switch.  (We've got things really
> stripped down trying to pin down the latency problem, so practically
> everything but our critical interrupt is commented out.)
>
> Is 78us of latency for maskable interrupts a reasonable goal, or is

> that outside of RTEMS' performance envelope?  Any comments are welcome.

FYI:

On an MVME 2306 board (300MHz PPC 604) I measured an _average_
interrupt latency better than 3us. The maximal latencies (over a couple of
hours
of operation) were as follows:

 - better than ~ 25us idle or under network load (flood pinging the MVME)
 - better than ~ 45us under heavy network and serial console load
                                  (flood ping to MVME, netcat from chargen
port on host
                                   directed to MVME's serial console).

I have submitted a paper to the upcoming ICALEPCS 2001 conference

http://arXiv.org/abs/cs.OS/0111035

comparing RTEMS to RTLinux and vxWorks (focussing on latencies). The
bottom line is: RTEMS performance is comparable to vxWorks, RTLinux
is much worse under load.

-- Till.