On Feb 12, 2009, at 14:01 , David Farrell wrote: > OK - I have been working with version 4.0.4 of pw that has > been compiled with only the internal espresso libraries (just > to check that it isn't an IBM library thing). The errors still appear.
the same example works on the cray with internal libraries as well. > I am trying my 'simple' periodic 1152 atom cell try first to reproduce the example I sent. There is no way you can run large jobs without distributing the subspace matrices. Paolo --- Paolo Giannozzi, Democritos and University of Udine, Italy
