On Thu, 2012-08-23 at 18:34 +0200, Eduardo Ariel Menendez Proupin wrote: > I generated the file vdW_kernel_table with version 5.0.1 and it is different > from the file coming with the 4.3.2. Should they be different ?
they shouldn't, but apparently the kernel is given by an integral that is numerically unstable at small values of q, so you may get slightly different numbers with different machines/compilers/phases of the moon. This shouldn't however affect the final results: those large numbers you get are multiplied by very small densities. P. -- Paolo Giannozzi, IOM-Democritos and University of Udine, Italy