Le 2 November 2006, Steven G. Johnson, à bout, prit son clavier pour taper sur son écran: > On Thu, 2 Nov 2006, Loïc Le Guyader wrote: > >(define-param dt 0.5) > >[...] > >(set! resolution 10) > >(run-until 60 (to-appended "ex" (at-every dt output-efield-x))) > >[...] > >But with dt=0.1 for example, I don't get 60*.1=600 data points: > >meep dt=0.1 test.ctl ; h5ls test-ex.h5 > > I think you are just seeing rounding error. > > With a resolution of 10, the time-step is 0.05. So, with dt = 0.1 it
It doesn't depend on the resolution. I tried with resolution up to 100 and I still get 507 points. > Right now, it's more susceptible to rounding than it has to be because > Meep waits to output until the time is >= last-output-time + dt. It would > be better to output as soon as the time is within half a time-step of the > requested output time. A patch to do this is: > > hunk ./libctl/meep.scm.in 442 > - (if (or (eq? to-do 'finish) (>= T (+ Tlast dT))) > + (if (or (eq? to-do 'finish) $ > + (>= T (+ Tlast dT (* -0.5 (meep-fields-dt-get fields))))) > > With this patch, Meep should give you your expected 600 data points. And it works ! Excellent ! \o/ Cheers. -- _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
