|
[Sponsors] |
January 21, 2009, 06:13 |
GPU and LBM
|
#1 |
Guest
Posts: n/a
|
Has anyone tried to program the Lattice Boltzmann method using a GPU library?
|
|
January 21, 2009, 07:01 |
Re: GPU and LBM
|
#2 |
Guest
Posts: n/a
|
I haven't but people who program computer games have
http://www.cg.tuwien.ac.at/research/...er_2008_FluSi/ |
|
January 21, 2009, 13:07 |
Re: GPU and LBM
|
#3 |
Guest
Posts: n/a
|
Soon we will use crysis engine to simulate CFD...
|
|
January 22, 2009, 04:59 |
Re: GPU and LBM
|
#4 |
Guest
Posts: n/a
|
||
January 22, 2009, 06:19 |
Re: GPU and LBM
|
#5 |
Guest
Posts: n/a
|
That would be a waste of resources.
|
|
January 22, 2009, 07:44 |
Re: GPU and LBM
|
#6 |
Guest
Posts: n/a
|
There's always this, an even greater waste of resources
http://www3.interscience.wiley.com/j...39484/abstract (not LBM though) |
|
January 23, 2009, 07:56 |
Re: GPU and LBM
|
#7 |
Guest
Posts: n/a
|
Well, I'm not an expert in the Lattice Boltzmann method but it it requires a huge amount of data to manipulate, it is never a good idea to use GPU. GPU can hold very limited amount of data itself and so, usually, its speed gets very slow compared to a fast CPU.
However, the data to manipulate is small, e.g. breaking some encryption, it GPU is very fast and works excellent. I hope u can decide now if u should choose GPU or CPU for ur computation. |
|
February 15, 2009, 09:31 |
Re: GPU and LBM
|
#8 |
Guest
Posts: n/a
|
I disagree! Check out papers 22 and 23 on this page: http://www.irmb.bau.tu-bs.de/institu...-tolke_de.html
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Will GPU's play a part in the future of CFD? | Tom | Main CFD Forum | 15 | January 22, 2009 05:04 |
Mixed CPU/GPU computing | Joe | Main CFD Forum | 4 | September 20, 2006 12:04 |