[Paraview-developers] An Abstract withs some paraview in it

[Biddiscombe, John A.]

Guys,

Apologies to anyone who may be on this list who I've already sent it to for duplication.

Title : "In-Situ Visualization and Analysis of SPH Data using a ParaView Plugin and a Distributed Shared Memory Interface"

It's a bit off topic, but since I see so much co-processing activity going on (which looks excellent by the way, and I'll soon be testing it out) - and we have our own framework basically running quite well and stable now. I thought you might like to see an abstract we sent off about a week ago to an SPH conference (in the UK in June this year).  (it's the main place where pv-meshless gets shown off).

Thought you might be interested in the progress we've made.
a couple of things not mentioned in the abstract which make it possibly useful (though if I'd known about the coprocessng stuff before we wrote our proposal, I may not have gone down this path!)


1)      Great if simulation runs on a bare bones machine, and you want analysis/vis on a GPU cluster and have some GPU enabled filters in paraview (we don't  yet)

2)      Great if memory is a problem on simulation box, and you don't want to link in and visit/paraview stuff.

3)      Great if the simulation scales really well and you can use thousands of cores, but your analysis scales badly and needs a small number of Fat nodes rather than tons of lean fast ones.

ftp://ftp.cscs.ch/out/biddisco/temp/abstract.pdf

JB


--
John Biddiscombe,                            email:biddisco @ cscs.ch
http://www.cscs.ch/
CSCS, Swiss National Supercomputing Centre  | Tel:  +41 (91) 610.82.07
Via Cantonale, 6928 Manno, Switzerland      | Fax:  +41 (91) 610.82.82

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview-developers/attachments/20100312/35f17790/attachment.htm>