[Olsr-dev] Improving SPF with binary heaps

[Diogo Gonçalves]

Hi,

I looked at the bugged function on olsrd2 code but I don't realized the bug
there.

I will start the performance tests but I still have a problem with CORE.

Olsrd2 seems run but nothing is printed at terminal. Is it OK? Is there a
parameter to enable the debug mode?

I read something about logging here [1] but it did not solved my problem.

Best regards!

[1] http://olsr-dev.olsr.narkive.com/lvplqXIu/how-to-enable-logging-in-olsr2

2015-10-08 5:56 GMT-03:00 Henning Rogge <(spam-protected)>:

> I just found the bug and fixed it, it was not related to the heap code...
>
> Can you maybe try to see if you can still find a performance
> difference between the "heap" branch and the "master" branch of
> OLSRv2?
>
> I would be interested if the heap makes a good difference.
>
> Henning
>
> On Tue, Oct 6, 2015 at 5:05 AM, Diogo Gonçalves
> <(spam-protected)> wrote:
> > ok, I will work on it.
> >
> > 2015-10-05 5:38 GMT-03:00 Henning Rogge <(spam-protected)>:
> >>
> >> Yes...
> >>
> >> I tested the olsrd2 "heap" branch a bit more and run into lots of
> >> segfaults. They don't happen in the master branch, so the problem
> >> should either be in the modified routing code or the the heap code.
> >>
> >> Olsrd2 should run without modifications in CORE... just run
> >> "olsrd2_static <interface1> <interface2> ..." in each node.
> >>
> >> I can see the segfault in a small grid topology at the moment, but I
> >> have trouble tracking the problem down.
> >>
> >> Henning
> >>
> >> On Mon, Oct 5, 2015 at 10:36 AM, Diogo Gonçalves
> >> <(spam-protected)> wrote:
> >> > Hi,
> >> >
> >> > I looked at the branch and the code looks fine.
> >> >
> >> > Was intentional the single @param comment at heap_is_node_added
> >> > function?
> >> > Apart from that the code looks great.
> >> >
> >> > Can I help you with something? I still have been trying to run olsrd
> on
> >> > CORE
> >> > but without success, when I run my olsrd on core I will send you the
> >> > performance report.
> >> >
> >> > Best regards.
> >> >
> >> >
> >> >
> >> > 2015-10-04 12:42 GMT-03:00 Henning Rogge <(spam-protected)>:
> >> >>
> >> >> Hi,
> >> >>
> >> >> I just added the heap code to the OLSRv2 repository (in a branch
> >> >> called "heap") and did a few basic tests...
> >> >>
> >> >> I have not (at the moment) added the heap code to the "src/common"
> >> >> directory because I still have to rework it for generic "key"
> >> >> variables.
> >> >>
> >> >> Feel free to have a look at the branch.
> >> >>
> >> >> Henning
> >> >>
> >> >> On Fri, Sep 18, 2015 at 11:23 AM, Henning Rogge <(spam-protected)>
> >> >> wrote:
> >> >> > We had a fix for olsrd some time ago which was important for
> Core...
> >> >> >
> >> >> > because Core forgets to set a good broadcast address.
> >> >> >
> >> >> > Merge the latest olsr master branch into your binary heap branch
> and
> >> >> > try
> >> >> > again.
> >> >> >
> >> >> > Henning
> >> >> >
> >> >> > On Thu, Sep 17, 2015 at 11:21 PM, Diogo Gonçalves
> >> >> > <(spam-protected)> wrote:
> >> >> >> Hi,
> >> >> >>
> >> >> >> Sorry for my long time offline, I had some problems at home that
> >> >> >> disturbed
> >> >> >> me.
> >> >> >>
> >> >> >> I finish my new patch [1] following the plain, if i did something
> >> >> >> wrong
> >> >> >> about the plain, tell me and i'll fix it.
> >> >> >>
> >> >> >> Henning suggested me to use CORE to test my code, I installed the
> >> >> >> emulator
> >> >> >> but I haven't found a way to configure CORE to run my olsrd. Is
> >> >> >> there a
> >> >> >> reference to learn more about olsrd on CORE?
> >> >> >>
> >> >> >> Again, sorry for my time out. I'm available to work on this
> project
> >> >> >> and
> >> >> >> I
> >> >> >> want to do this.
> >> >> >>
> >> >> >> Best regards
> >> >> >>
> >> >> >>
> >> >> >>
> >> >> >> [1] https://github.com/diogomg/olsrd-binary-heap
> >> >> >>
> >> >> >> 2015-08-21 1:40 GMT-03:00 Henning Rogge <(spam-protected)>:
> >> >> >>>
> >> >> >>> On Fri, Aug 21, 2015 at 2:30 AM, Diogo Gonçalves
> >> >> >>> <(spam-protected)> wrote:
> >> >> >>>>
> >> >> >>>> Hi,
> >> >> >>>>
> >> >> >>>> I'm finishing my patch, I'll send it to you soon.
> >> >> >>>>
> >> >> >>>> For now, I just have some issues about how to measure the
> >> >> >>>> performance
> >> >> >>>> of
> >> >> >>>> the priority queues in olsrd. I'm using olsr switch to test my
> >> >> >>>> code
> >> >> >>>> and I
> >> >> >>>> did not find a good way to scale my network and measure the
> >> >> >>>> performance. I
> >> >> >>>> have some results outside the olsrd(see attachments) using
> >> >> >>>> kcachegrind and
> >> >> >>>> other tools to do this and now I'm trying to measure the
> >> >> >>>> performance
> >> >> >>>> into
> >> >> >>>> the olsrd too.
> >> >> >>>>
> >> >> >>>> Is olsr switch the best way to do this?
> >> >> >>>
> >> >> >>>
> >> >> >>> No, it is definitely not.
> >> >> >>>
> >> >> >>> I would go with an emulator like CORE (from the NRL) or maybe
> setup
> >> >> >>> a
> >> >> >>> bunch of VMs on my own.
> >> >> >>>
> >> >> >>> Henning
> >> >> >>>
> >> >> >>
> >> >
> >> >
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.olsr.org/pipermail/olsr-dev/attachments/20151009/ce99dd68/attachment.html>