[lkml]   [2001]   [Sep]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH] Preemption Latency Measurement Tool
Andre Pang wrote:
> On Sat, Sep 22, 2001 at 02:10:18AM -0400, Robert Love wrote:
> > > i did a test of it on linux-2.4.10-pre13 with Benno Senoner's
> > > lowlatency program, which i hacked up a bit to output
> > > /proc/latencytimes after each of the graphs. test results are at
> > >
> > >
> > >
> > > and since i stared at the results in disbelief, i won't even try
> > > to guess what's going on :). maybe you can make some sense of
> > > it?
> >
> > Well, its not hard to decipher...and really, its actually fairly good.
> > the latency test program is giving you a max latency of around 12ms in
> > each test, which is OK.
> arrgh! i just realised my script buggered up and was producing the same
> graph for all the results. please have a look at the page again, sorry.
> apart from that, i'm still confused. compared to other graphs produced
> by the latencytest program, my system seems to have huge latencies.
> unless i'm reading it wrongly, the graph is saying that i'm getting
> latencies of up to 30ms, and a lot of overruns. compare this to
> which shows latencytest on 2.4.0-test2, and
> which are the results for latencytest on 2.2.10. admittedly these
> kernels are much older, but i'm consistently getting far more latency
> than those kernels. that's the bit i'm confused about :) i've tried
> Andrew Morton's low-latency patches as well, to no avail. i've made
> sure i've tuned my hard disks correctly, and i don't have any other
> realtime processes running.
> am i concerned with a different issue than the one you're addressing?
> > the preemption-test patch is showing _MAX_ latencies of 0.8ms through
> > 12ms. this is fine, too.
> yep, i agree with that ... so why is latencytest showing scheduling
> latencies of > 30ms? i get the feeling i'm confusing two different
> issues here. from what i understand, /proc/latencytimes shows the
> how long it takes for various functions in the kernel to finish, and
> the latencytest result shows how long it takes for it to be
> re-scheduled (represented by the white line on the graph).

The one thing the latancytimes patch doesn't monitor is interrupt off
time. Maybe it should...

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:03    [W:0.142 / U:0.096 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site