[FLASH-USERS] potential solution to refinement problem

Andrea Gatto andreag at MPA-Garching.MPG.DE
Sun Feb 15 10:59:13 EST 2015


Dear all,

concerning the problem

http://flash.uchicago.edu/pipermail/flash-users/2014-May/001456.html
http://flash.uchicago.edu/pipermail/flash-users/2014-May/001458.html

a solution has been found in switching back to pure injection instead of
quadratic interpolation between blocks by setting

interpol_order = 0

I can't be 100% sure that the problem won't show later on in the
simulation since high resolution simulations are expensive to run, however
this fixed the problem for 3 setups with different physics included plus a
different setup with another version of the code.

Linear interpolation (interpol_order = 1) gives the same problem but
slightly later.

Best,

Andrea




More information about the flash-users mailing list