using a negative dielectric with kspace PPPM

Hi all,

I’m trying to simulate gravity by setting "Dielectric -1” and charge q= sqrt(G/k)*m as in the paper:

“Substellar fragmentation in self gravitating fluids with a major phase transition”
Füglistaler and Pfenniger, A&A 2015

https://www.aanda.org/articles/aa/pdf/2015/06/aa24798-14.pdf

However, when I set the dielectric constant to a negative value, I get the error “Could not compute g_ewald,” indicating that the Newton-Raphson solver in PPPM::adjust_gewald() fails to converge. Unfortunately I’m not quite savvy enough with the PPPM solver to easily figure out what is causing this to break. Does anybody have any insight?

Thanks,

Hi all,

I’m trying to simulate gravity by setting "Dielectric -1” and charge q= sqrt(G/k)*m as in the paper:

“Substellar fragmentation in self gravitating fluids with a major phase transition”
Füglistaler and Pfenniger, A&A 2015

https://www.aanda.org/articles/aa/pdf/2015/06/aa24798-14.pdf

However, when I set the dielectric constant to a negative value, I get the error “Could not compute g_ewald,” indicating that the Newton-Raphson solver in PPPM::adjust_gewald() fails to converge. Unfortunately I’m not quite savvy enough with the PPPM solver to easily figure out what is causing this to break. Does anybody have any insight?

this is just a guess, but how about doing a “run 0” with the unmodified dielectric, take note of grid and g_ewald settings, then set them manually and try with dielectric -1 again?

axel.