[ltt-dev] tracepoint layer and marker layer

[Jiaying Zhang]

On Fri, Sep 19, 2008 at 11:23 AM, KOSAKI Motohiro <
kosaki.motohiro at jp.fujitsu.com> wrote:

> Hi
>
> > Hi Folks,
> >
> > I have a general question about LTTng kernel code. As I understand, LTTng
> > adds an extra tracepoint layer between the traced subsystems and the
> > marker layer. The extra code added by the tracepoint layer seems to
> > introduce
> > noticeable overhead according to my measurements.
>
> Could you post your mesurement way and result?


I measured my results on a multi-processor machine with the tbench
benchmark http://samba.org/ftp/tridge/dbench/.
The detailed results were posted at
http://lists.casi.polymtl.ca/pipermail/ltt-dev/2008-September/000153.html
(I should have referred to the link in my original email, sorry about that).


> My past mesurement indicate tracepoint overhead is ignoreble low.
>

Could you share your measurements and results? E.g., the kernel version
and LTTng patch version, the benchmark you used, your LTTng kernel
configurations,
the machine type, and so on. Thanks a lot!

Jiaying

>
>
> > I wonder whether it will
> > be more efficient and simpler if we let subsystems call the marker layer
> > directly?
> > Any thought?
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.casi.polymtl.ca/pipermail/lttng-dev/attachments/20080919/77d81a3c/attachment-0002.htm>