the metal-GNRs-metal device calculation issues

[njuxyh]

 Hi all:

i plan to try to optimze the geometry of the Ni-GNRs(finit wider)-Ni device,

i have read the Ni-graphene interface manual, but i know the GNRs is different from graphene,

if i want to construct such a device, in order to screen the interaction of the mirror of neighbor GNRs, so the electrode size in the x-y plane should huge, as least 10 angstrom x direction, and in y direction should 10 angstrom+the length of GNRs, so the vacuum space for the GNRs is enough.
but in summary, the total number of atoms in the device is thousands including the atoms in both electrode and in the central region,

therefore, i am wondering such big device, is there any menthod to reduce the computational quantities? because now, my computation is stopped at The step:



|                                                                              |
| Left Electrode Calculation  [Started Wed May 14 16:06:47 2014]               |
|                                                                              |
+------------------------------------------------------------------------------+

+------------------------------------------------------------------------------+
| Checkpoint Handler                                                           |
| Filename : /tmp/checkpoint75331660.nc                                        |
| Interval : 0.5 h                                                             |
+------------------------------------------------------------------------------+

                            |--------------------------------------------------|
Calculating Eigenvalues    :




anyone has simulate such system?

i need your help

 

[GJK]

Hi njuxyh,

I have made and tried to optimize the system using using GNR with Au (Gold) electrode.But my system had around 500 atoms.My optimization failed saying Buzz error.I tried three times but the same Buzz error.

Dear all,
Kindly help us to resolve this issue.Our system is ATK 12.8.But we cannot proceed for optimization for atoms >150(number of atoms).So we are constraining our work.I have tried with even 165 atoms but optimization stopped.But for the same system with atoms<150 I got results.Kindly clarify

Thanks in advance

[Anders Blom]

ATK 2014 will have methods for reducing the memory consumption of in particular device calculations dramatically. But there is no way around the fact that if you want to simulate 500 atoms, you need a big computer.

[GJK]

ATK 2014 will have methods for reducing the memory consumption of in particular device calculations dramatically. But there is no way around the fact that if you want to simulate 500 atoms, you need a big computer.

Dear Mr.Anders,

So you say that if i want to simulate 500 atoms means I need to purchase a big computer.That is what you mean am I right??

Thanks in advance

[Anders Blom]

500 atoms with a classical potential, no that you can do on anything. But 500 atoms DFT is serious business.

[njuxyh]

this means :
if i want to simulate a device with atoms about 500 ,  i should have big compute and i can only use the semi-classic method not a DFT method?

and how is the big compute?

how is the memory is how? how is the cpu ?

thanks very much!