Dear Stephen, Thanks for this advice - sadly I don't have much choice - I need to simulate a conducting fluid and that means using a drude like model in which the omega term should be zero. Since the freqeuency is 3.3e-7 I need something smaller than this by a decent factor and have tried 1e-20 and 1e-10 and 1e-9 (with appropriate shanges to gamma and delta_epsilon) with no improvement so far. Running for just one step I get an output which is all NaN so the supposition of overflows is correct. It's quite a strange case I guess - the vacuum wavelength for the frequency I'm using is 3000km so the scaling would probably prefer me to use this kind of number. In seawater of course the wavelength comes out more like 154m (due to the effect of the conductivity on the magnitude of the propogation constant). Sadly I need to see what happens in the 10-300m range and thus there's a problem foe me to set the scale unit to something approximating the wavelength and still get enough points calculated.... ARgh! I guess that there are some situations where SI units are rather useful after all.
unless there's any major thought out there I think this has me stumped all the best Chris -------- Original Message -------- Subject: Re: [Meep-discuss] Very slow computation for complex dielectrics Date: Thu, 2 Nov 2006 14:32:12 -0500 (EST) From: Steven G. Johnson <[EMAIL PROTECTED]> Reply-To: [EMAIL PROTECTED] To: meep-discuss <[email protected]> References: <[EMAIL PROTECTED]> On Tue, 31 Oct 2006, Dr Chris Stevens wrote: > am still playing with this complex dielectric issue. I am running > two ctl files listed below. The first, sea_fake.ctl runs in about 75s > having only real dielectrics ie water with zero conductiivty. The second > which simulates the same situation but with complex dielectric set by > using a seawater conductivitty of 4S/m takes over 200000s to run! > I'm using meep which responds to the meep -V command with: Probably something is blowing up, and so you're seeing the CPU slow down because of floating-point exceptions. > (polarizations > (make polarizability > (omega 1e-20) (gamma 1e13) (delta-epsilon 4.5e65)) > ) > ))) Try to use smaller values for these things, e.g. put resonances not at omega=0. I've noticed that trying to put a resonance at omega near zero like this can lead to instabilities. _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss -- .................................................................. . Dr.Chris Stevens . Oxford University . . Tel: +44 1865 283272 . Department of Engineering Science . . Fax: +44 1865 273905 . Parks ROad . . [EMAIL PROTECTED] . Oxford, OX1 3PJ, UK . .................................................................. _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
