Re: 2.6.14-rt1

[Carlos Antunes <cmantunes@gmail.com>]

On 11/1/05, Steven Rostedt <rostedt@goodmis.org> wrote:
> On Tue, 2005-11-01 at 22:26 -0500, Carlos Antunes wrote:
>
> >
> > It's a simple program I put together to test wakeup latency. Each
> > thread basically sleeps for 20ms, wakes up and executes a couple of
> > instructions and goes back to sleep for another 20ms. Multiply this by
> > a thousand. What I found out is that, inthis situation, and using
> > realtime-preempt, SCHED_OTHER offers 3 orders of magnitude less
> > latency than SCHED_FIFO or SCHED_RR. Which suggests to me there is
> > something fishy going on.
>
> Could you supply this program?  I like to see what it does on my
> systems.
>

Steve,

Here's the thing:
http://www.nowthor.com/OpenPBX/timing.c

Let me know what kind of results you get.

Thanks!

Carlos


--
"We hold [...] that all men are created equal; that they are
endowed [...] with certain inalienable rights; that among
these are life, liberty, and the pursuit of happiness"
        -- Thomas Jefferson