This thread is locked.Only browsing is available.
Top Page > Browsing
Transmission
Date: 2011/05/27 21:42
Name: lida   <lida.ansari@tyndall.ie>

Dear openmx developers,

I calculated successfully the NEGF calculation (step2 in transmission calculations) for a big system. However, it ran out of memory using the same number of nodes for the third step. I wonder if it is normal, since the code read in the Hamiltonian from the NEGF calculation (step 2).

I would appreciate if you could give me any advice to solve this problem.

Regards,
Lida
メンテ
Page: [1]

Re: Transmission ( No.1 )
Date: 2011/05/28 10:05
Name: T.Ozaki

Hi,

One practical solution could be to use only few (one) cores in a node consisting of
multiple processors in the parallel calculation so that the memory space per core can
be enlarged.

Regards,

TO
メンテ

Page: [1]