parallelization problem in rfstrs
Posted: Fri May 24, 2013 5:19 pm
Hi,
I am doing a calculation of strain response, in a tetragonal system with 192 k points. I find that the rfstrs computation works on 32 processors for longitudinal strain in directions 1 and 2 but not 3 (it hangs), while on 24 processors all directions work. Note that both 32 and 24 are factors of 192. Any ideas where this would be failing?
Joe
I am doing a calculation of strain response, in a tetragonal system with 192 k points. I find that the rfstrs computation works on 32 processors for longitudinal strain in directions 1 and 2 but not 3 (it hangs), while on 24 processors all directions work. Note that both 32 and 24 are factors of 192. Any ideas where this would be failing?
Joe