Hi, Steven and meep users:
I am evaluating the memory I need and read related discussion in
http://osdir.com/ml/science.electromagnetism.meep.general/2006-09/msg00054.html
So based on that discussion, assuming the total grid points are 10^6, in
3D structure.
And, in FDTD, it has to handle 6 fields, ex,ey,ez, hx, hy, epsilon and hz.
Each needs 3 nodes, x, y,z for 3D structure.
So total nodes we need are 10^6*7*3=21*10^6. Each node needs double value,
does it mean we need 159.6Mbytes, at least?
And we have to multiple a 3 by this number? but I don't know why.
Also, I used to handling dispersive material. In that discussion, we have
to handle extra 6 fields, I think those are Dx, Dy, Dz, Px, Py, and Pz
due to time-convolution demand for Drude-Loreztian model.
So total memory we need, prior to multiple by 3, will become
10^6*(7+6)*3*7.6=296Mbytes. is this correct?
Is there any reason to increase the memory usage except above unavoidably?
Thank you in advance
Best Regards
Adrian
______________________________________________________________________________________
怎樣減少垃圾信？只要看到垃圾信，立即按下「這是垃圾信」按鈕。http://tw.promo.yahoo.com/antispam/index.html_______________________________________________
meep-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
