[Pw_forum] wrong results from parallel running of ph.x

[Paolo Giannozzi]

On Thu, 2012-10-11 at 16:02 +0200, Jia Zhang wrote:

> It is found that the parallel result seems incorrect 

which parallel results, obtained EXACTLY how, with which version 
of the code?


> And also, there is another grid method to run ph.x. If we have parallel 
> version and split the q points automatic why do we need the grid? 

you don't "need", you "may use" the grid (i.e. a system of
loosely coupled computers) if you have access to one.

P.
-- 
Paolo Giannozzi, IOM-Democritos and University of Udine, Italy