[BonnMotion] In Range Printer output times exceed duration
Alexander Bothe
bothe at uos.de
Wed Jan 31 18:34:36 CET 2024
Hi Tom,
I finally had some spare time to test this myself, using a fresh install
from the latest release (3.0.1) and the parameters from your .params file.
On my PC (AMD 7800X3D, 32GB RAM, Win10, JDK21), trace generation takes
around 1 second, the IRP calculation using the interval mode, even with
very small intervals seems to be fine as well.
Radius based calculations with a low radius also seem to be fine, issues
start to appear with a higher radius setting. I would assume the bug is
somewhere in the TopologyChangeTimes class [1]. Unfortunately, I
probably won't have time to hunt for it anytime soon, especially since I
plan to remove apps with the next BonnMotion major release (no ETA).
They probably will make a comeback in a separate project some day.
Nevertheless, could you please provide the parameters you are using for
your IRP calculations for some further testing?
Kind regards
Alexander
[1]
https://github.com/sys-uos/BonnMotion/blob/master/src/edu/bonn/cs/iv/bonnmotion/apps/helper/TopologyChangeTimes.java
On 08.12.23 10:38, Thomas Jordan wrote:
> Hi,
> I have recently started generating RandomWaypoint scenarios using
> BonnMotion and then using the In Range Printer application to analyse
> the results and I now have two questions.
>
> First question: I notice that when I run the In Range Printer on a
> scenario, maximum times in the output of the IRP exceed the duration of
> the scenario. For example, if I run a scenario for 1500, I see max
> waypoint times in the movements.gz of 1500.0 which looks correct, but
> see max times in the .irp file of 5076.121483475256.
>
> Can anyone explain to me why this is happening?
>
> The .param file for this scenario is as follows:
>
> model=RandomWaypoint
>
> ignore=3600.0
>
> randomSeed=3
>
> x=500.0
>
> y=500.0
>
> duration=1500.0
>
> nn=64
>
> circular=false
>
> J=2D
>
> dim=4
>
> minspeed=2.0
>
> maxspeed=4.0
>
> maxpause=0.0
>
> Second question: the generation of the IRP file from a scenario seems to
> take a long time. For the above scenario, I see times 2.5 – 3 hrs on an
> Intel(R) Core(TM) i7-9700 CPU @ 3.00GHz with 32 GB of RAM (Windows 11,
> Java v21.0.1). Is this apparently long run-time for IRP herer simply
> because of the computational time taken to do the physics/motion
> calculations (which I assume is some calculation of moving point
> colliding with moving circle and emerging from the circle while both are
> still moving), or is this something I can do to reduce it?
>
> Regards,
>
> Tom
>
> *Thomas Jordan*
>
> Principle Software Engineer*
>
> **Multiple Access Communications Limited*
>
> office: +44 (0)23 8076 7808
>
> email: thomas.jordan at macltd.com
>
> web: www.macltd.com <http://www.macltd.com>
>
>
> _______________________________________________
> BonnMotion mailing list
> BonnMotion at list.serv.Uni-Osnabrueck.de
> https://list.serv.uni-osnabrueck.de/mailman/listinfo/bonnmotion
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 6039 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://list.serv.uni-osnabrueck.de/pipermail/bonnmotion/attachments/20240131/20279981/attachment.p7s>
More information about the BonnMotion
mailing list