[Olsr-dev] [OLSRD energy consumption ]

[Lakatos-Tóth Pál]

2010. október 5. 14:04 Henning Rogge írta, <(spam-protected)
>:

> On Tue October 5 2010 13:32:42 Lakatos-Tóth Pál wrote:
> > Dear OLSRD Developers,
> >
> > I write this letter, because I want to ask for advice in the topic of
> OLSR
> > Energy consumption measurement. We develop a framework, which is
> simulating
> > an ARM platform and can run a virtual linux, where the olsrd is running.
> We
> > can simulate some nodes, which can communicate with each other and we
> would
> > like to measure the energy consumption of OLSRD. We have a plan about
> > measuring , but we need to insert special instructions to the source code
> >  of the daemon, which signals the entry and exit points of these
> high-level
> > functional blocks, because we would like to analyze deeply what is
>  really
> > happen inside of OLSRD. If it is possible we would like to ask for some
> > hints about the interesting parts of the protocol. Since the developers
> are
> > the most familiar with the code it could be interesting to gather some
> > useful information where to place these labels. Basically this labeling
> > will create a high level infromation mapping with the measured energy
> > values, so we can directly read from diagrams what was being done during
> > an energy consumption change.
>


> You could use a profiler like valgrind to measure how much CPU time was
> spend
> in each function of OLSRd.
>
> Henning Rogge
> --
> Diplom-Informatiker Henning Rogge , Fraunhofer-Institut für
> Kommunikation, Informationsverarbeitung und Ergonomie FKIE
> Kommunikationssysteme (KOM)
> Neuenahrer Straße 20, 53343 Wachtberg, Germany
> Telefon +49 228 9435-961,   Fax +49 228 9435 685
> mailto:(spam-protected) http://www.fkie.fraunhofer.de
> GPG: E1C6 0914 490B 3909 D944 F80D 4487 C67C 55EC CFE0
>

Dear Henning,



Thank you for your reply, but I think you do not understand me, what I would
like to get if it possible. So we have a framework, which is in developement
phase and we can debug the olsrd in very low level with the help of our ARM
Simulator ( it is better and lower level than valgrind ). For example we
place a label to the source where the OLSR started to calculate the shortest
path between nodes. After we recompile the olsrd with this label at the
start and the end of this function and after we run that binary on our
energy measurement framework. After it run,  we can see on the output the
energy consumption chart and we have a part of a graph which is indicated
with the signaled function's energy. So we can see that the energy
consumption increased by (for example) 0.05 Joule/sec, so we have an
information about how much energy spent to calculate the shortest path. So I
would like to get not the all, but some advice from the developers about
where I would place these signals to the source and what are the most
interested parts of the OLSRD in aspect of energy consumption
measuring,  because
I think they have more information about how is the olsrd built up, so they
can advise some useful information if it possible .

Thanks in advance,



Pal Lakatos-Toth

-- 

Lakatos-Tóth Pál
Software Engineer
Department of Software Engineering
University of Szeged
SZEGED, H-6725
Honved ter 6
HUNGARY
Tel: +36 (62) 54-6783
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.olsr.org/pipermail/olsr-dev/attachments/20101006/82466045/attachment.html>