yade-users team mailing list archive

[Václav Šmilauer <eudoxos@xxxxxxxx>]

> Presently, I am very 'lost'... I have just started DEM.  YADE is my
> first DEM software.    I've just started my Phd and the topic is open
> (cuz i'm not doing sth on a funded project)... In fact, i've not
> decided what project to do yet.  i have to consider also the
> 'technical' difficulties when deciding a topic.  The community is my
> only source of help.   Some things are v difficult and may have to be
> abandoned, e.g. the algorithm to generate rock blocks from joint
> planes (like 3Dec, not PFC).
>  
> I was looking at ellipsoids..On google, found these websites:
> http://local.wasp.uwa.edu.au/~pbourke/miscellaneous/implicitsurf/
> http://www.frank-buss.de/lisp/polygonizer.html
> When I come to think of it (still thinking), if i use some weird
> shapes to do simulation, and do some experimental calibration = PhD
> dissertation?  I think i'll have to finalize what i'll be doing for my
> phd soon <headache>...

Most the time you will adjust the topic as you progress, so unless your
supervisor is very strict, you can try one thing which will give you
bigger picture and the you will know better.

If you are into computing and programming, then exploring/implementing
high-performance algorithms: 2 things coming to mind are 1. domain
decomposition and distributed memory parallelism (MPI), 2. DEM on
OpenCL/CUDA chips. Both of them might be very nice topic for a PhD, the
first would entail work on yade itself, the second probably a separate
code. DEM on OpenCL is a work that's probably never been done (OpenCL is
no older than 1 year), if that matters.

Cheers, Vaclav