Hello, I'm interested in getting some integrated quantities from a cylindrical 2d FLASH simulation. Just to be sure, I started by comparing the total simulation mass computed by yt ( with pf.h.all_data().quantities['TotalQuantity'] etc.) and the one provided in FLASH logs, and the results are rather different [see attachment]: * _Values_: difference of up 8% between the two methods * _Time evolution_: total mass is decreasing in time according to FLASH (which is ok with open boundary conditions) but it is increasing with the yt's integration. * _Sensitivity to resolution_: there are jumps in total density given by yt for the time steps where the maximum refinement level is manually decreased. Would anyone know what might be the reason for this behaviour? Just in case, I tried to do a second order integration instead of simply summing the values in all the cells, but if didn't make an error, the correction was negligible. I can provide data and the yt script I'm using by PM if some could look into this. Thank you, Regards, -- Roman Yurchak