[mcstas-users] Source Parallelization

[Federico M. Bertalot]

Hello,

i have written my own source component, and in order to normalize the 
results i defined the weight of each neutron as:

p=1.0/mcget_ncount()

it works fine for one processor, but for more processors mcget_ncount() 
seems to return the number of neutron histories running on each 
processor, instead of the total number of neutron histories. Then, the 
normalization of the results changes with the number of processors. 
Which is the best solution to this problem?


Thank you very much,

Federico.