Dear Francesco, Your situation is the one we had to face some months ago.
After some months of clustering experience I can tell you the following: all depends on the "size" of the structure you need to simulate. Let's understand by size the "computational size", i.e. how much memory you need for your structures. If that memory can be allocated in one single computer, I would recommend you to use a multi-core single computer, since you avoid the burden of a cluster administration. If not, go for a cluster solution, since the structure size is divided amongst the computational nodes. In the latter case it is also advisable to have multipro/core nodes. Some time ago we made an estimation of the size of the structures, you can find in: http://www.gco.upv.es/publis/pdf/domenech-epixnet-ws-2007.pdf I hope this helps, if you have more question, please ask! Pascual. Francesco Riboli escribió: > Dear all, > > we are planning to buy cluster of computers to perform > extensive numerical simulations. In order to choose the best system > for meep, we need to know for which type of architecture Meep is optimized. > It is better to have one node with multiprocessors and multicore > or to distribute the calculus on diverse node of the cluster? > Many Thanks > > Francesco > > ------------------------------------------------------------------------ > > _______________________________________________ > meep-discuss mailing list > [email protected] > http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss -- Dr. Pascual Muñoz Muñoz - Telecommunication Engineer Optical Communications Group - iTEAM Research Institute [EMAIL PROTECTED] - www.gco.upv.es - www.upv.es ------------------------------------------------------- Proud to be Valencian, http://www.valencian.org/ _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
