Messages

1

General Questions and Answers / Re: Reg. Geometric optimization in VNL

« on: September 3, 2009, 10:16 »

Thanks for the reply, zh.

I guess there is some confusion. So, I have both the co-ordinates in cartisan form (bulk-builder has option of showing the co-ordinates in both fractional as well as cartisan format). So, the cartisan-coordinates in my script file 'are matching' with those in the log-file. Since, I selected quasi-newton in the VNL, these should NOT match.

So, could you please re-iterate how do we do the geometrical-optimization using VNL. I mean, do I need to change something in the script-file also, although I have initially selected the method in VNL???

Thanks.

2

General Questions and Answers / Reg. Geometric optimization in VNL

« on: September 3, 2009, 09:27 »

Hi all,

I am curious to know how do I do geometric-relaxation for bulk when using VNL? Thing is that when I prepare the bulk structure and drop it on the nano-scripter, I can set the geometric-optimization method(either Quasi-newton or steepest-decline) and its relevant parameters. But when I start running the script, the first few lines in the job-manager window that shows the co-ordinates are not different from my initial co-ordinates. I would expect them to be different if the relaxation has taken place in the structure.

Thanks.

3

General Questions and Answers / Re: Urgent: Reg. molecular builder

« on: September 1, 2009, 23:31 »

Thanks zh. I think it should solve my problem now.

4

General Questions and Answers / Re: Urgent: Reg. molecular builder

« on: September 1, 2009, 06:22 »

Thanks Dr. Bloom. It works!! 

@zh: Could you please explain to me the first point you wrote.

Thankyou.

5

General Questions and Answers / Urgent: Reg. molecular builder

« on: August 28, 2009, 09:40 »

Hi everyone,

Is there some way/trick that we can put the distorted-CNT prepared using nanotube builder onto molecular builder?

Thanks.

6

General Questions and Answers / Re: Reg. spin-values for corresponding atoms in the interface

« on: August 20, 2009, 05:14 »

Thankyou ZH.
I actually did it that way. Thing is that after the first five Fe atoms, when we move to the second set, few of those actually belong to the right interface and few to the left interface. So, while assigning the spin in the anti-parallel case, I had to be very very cautious(and it was really tedious,) and hence this question for if there was any smarter way out for systems even larger than this.

Thanks.

7

General Questions and Answers / Reg. spin-values for corresponding atoms in the interface

« on: August 19, 2009, 23:04 »

I have a question regarding the initial-scaled spin values that we put in the central-scattering region. I prepare Fe-MgO-Fe two probe system, with 5Fe(left)-5MgO-4Fe(right) layers in the 'scattering region'.
with 2by2 repetition in AbyB direction.
The script file for this region shows as follows:

# Scattering elements and coordinates
scattering_elements = [Iron,      Iron,      Iron,      Iron,     
                       Iron,      Oxygen,    Oxygen,    Oxygen,   
                       Oxygen,    Oxygen,    Magnesium, Magnesium,
                       Magnesium, Magnesium, Magnesium, Iron,     
                       Iron,      Iron,      Iron,      Iron,     
                       Iron,      Iron,      Iron,      Iron,     
                       Oxygen,    Oxygen,    Oxygen,    Oxygen,   
                       Oxygen,    Magnesium, Magnesium, Magnesium,
                       Magnesium, Magnesium, Iron,      Iron,     
                       Iron,      Iron,      Iron,      Iron,     
                       Iron,      Iron,      Iron,      Oxygen,   
                       Oxygen,    Oxygen,    Oxygen,    Oxygen,   
                       Magnesium, Magnesium, Magnesium, Magnesium,
                       Magnesium, Iron,      Iron,      Iron,     
                       Iron,      Iron,      Iron,      Iron,     
                       Iron,      Iron,      Oxygen,    Oxygen,   
                       Oxygen,    Oxygen,    Oxygen,    Magnesium,
                       Magnesium, Magnesium, Magnesium, Magnesium,
                       Iron,      Iron,      Iron,      Iron]
scattering_coordinates = [[  0.7165 ,   0.7165 ,   1.433  ],
                          [  2.1495 ,   2.1495 ,   2.866  ],
                          [  0.7165 ,   0.7165 ,   4.299  ],
                          [  2.1495 ,   2.1495 ,   5.732  ],
                          [  0.7165 ,   0.7165 ,   7.165  ],
                          [  0.7165 ,   0.7165 ,   9.365  ],
                          [  2.1495 ,   2.1495 ,  11.56035],
                          [  0.7165 ,   0.7165 ,  13.7557 ],
                          [  2.1495 ,   2.1495 ,  15.95105],
                          [  0.7165 ,   0.7165 ,  18.1464 ],
                          [  2.1495 ,   2.1495 ,   9.365  ],
                          [  0.7165 ,   0.7165 ,  11.56035],
                          [  2.1495 ,   2.1495 ,  13.7557 ],
                          [  0.7165 ,   0.7165 ,  15.95105],
                          [  2.1495 ,   2.1495 ,  18.1464 ],
                          [  0.7165 ,   0.7165 ,  20.3464 ],
                          [  2.1495 ,   2.1495 ,  21.7794 ],
                          [  0.7165 ,   0.7165 ,  23.2124 ],
                          [  2.1495 ,   2.1495 ,  24.6454 ],
                          [  0.7165 ,   3.5825 ,   1.433  ],
                          [  2.1495 ,   5.0155 ,   2.866  ],
                          [  0.7165 ,   3.5825 ,   4.299  ],
                          [  2.1495 ,   5.0155 ,   5.732  ],
                          [  0.7165 ,   3.5825 ,   7.165  ],
                          [  0.7165 ,   3.5825 ,   9.365  ],
                          [  2.1495 ,   5.0155 ,  11.56035],
                          [  0.7165 ,   3.5825 ,  13.7557 ],
                          [  2.1495 ,   5.0155 ,  15.95105],
                          [  0.7165 ,   3.5825 ,  18.1464 ],
                          [  2.1495 ,   5.0155 ,   9.365  ],
                          [  0.7165 ,   3.5825 ,  11.56035],
                          [  2.1495 ,   5.0155 ,  13.7557 ],
                          [  0.7165 ,   3.5825 ,  15.95105],
                          [  2.1495 ,   5.0155 ,  18.1464 ],
                          [  0.7165 ,   3.5825 ,  20.3464 ],
                          [  2.1495 ,   5.0155 ,  21.7794 ],
                          [  0.7165 ,   3.5825 ,  23.2124 ],
                          [  2.1495 ,   5.0155 ,  24.6454 ],
                          [  3.5825 ,   0.7165 ,   1.433  ],
                          [  5.0155 ,   2.1495 ,   2.866  ],
                          [  3.5825 ,   0.7165 ,   4.299  ],
                          [  5.0155 ,   2.1495 ,   5.732  ],
                          [  3.5825 ,   0.7165 ,   7.165  ],
                          [  3.5825 ,   0.7165 ,   9.365  ],
                          [  5.0155 ,   2.1495 ,  11.56035],
                          [  3.5825 ,   0.7165 ,  13.7557 ],
                          [  5.0155 ,   2.1495 ,  15.95105],
                          [  3.5825 ,   0.7165 ,  18.1464 ],
                          [  5.0155 ,   2.1495 ,   9.365  ],
                          [  3.5825 ,   0.7165 ,  11.56035],
                          [  5.0155 ,   2.1495 ,  13.7557 ],
                          [  3.5825 ,   0.7165 ,  15.95105],
                          [  5.0155 ,   2.1495 ,  18.1464 ],
                          [  3.5825 ,   0.7165 ,  20.3464 ],
                          [  5.0155 ,   2.1495 ,  21.7794 ],
                          [  3.5825 ,   0.7165 ,  23.2124 ],
                          [  5.0155 ,   2.1495 ,  24.6454 ],
                          [  3.5825 ,   3.5825 ,   1.433  ],
                          [  5.0155 ,   5.0155 ,   2.866  ],
                          [  3.5825 ,   3.5825 ,   4.299  ],
                          [  5.0155 ,   5.0155 ,   5.732  ],
                          [  3.5825 ,   3.5825 ,   7.165  ],
                          [  3.5825 ,   3.5825 ,   9.365  ],
                          [  5.0155 ,   5.0155 ,  11.56035],
                          [  3.5825 ,   3.5825 ,  13.7557 ],
                          [  5.0155 ,   5.0155 ,  15.95105],
                          [  3.5825 ,   3.5825 ,  18.1464 ],
                          [  5.0155 ,   5.0155 ,   9.365  ],
                          [  3.5825 ,   3.5825 ,  11.56035],
                          [  5.0155 ,   5.0155 ,  13.7557 ],
                          [  3.5825 ,   3.5825 ,  15.95105],
                          [  5.0155 ,   5.0155 ,  18.1464 ],
                          [  3.5825 ,   3.5825 ,  20.3464 ],
                          [  5.0155 ,   5.0155 ,  21.7794 ],
                          [  3.5825 ,   3.5825 ,  23.2124 ],
                          [  5.0155 ,   5.0155 ,  24.6454 ]]*Angstrom

and the corresponding spins are written as follows: 

initial_scaled_spin = [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,
                            0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,
                            0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,
                            0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,
                            0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,
                            0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0. ]

Now, when I want to change the intial-scaled spin values for Fe atoms in the scattering region, I am confused if I should follow the structure and change the values in starting 20 and last 16 positions in the above given initial_scaled_spin matrix. OR, I should check where the Fe atoms from respective interfaces are placed in the scattering region and then change the spin-matrix correspondingly.(This is really confusing as there occurs a complete mix-up of Fe atoms from both interfaces while their position is being defined in the script file of VNL). 

Thanks.

8

General Questions and Answers / ATKError: Exceeded maximum number of self-consistent iterations.

« on: August 1, 2009, 09:08 »

I tried to follow exactly what the MTJ tutorial specifies(just to reproduce those results) and got the following result:

Traceback (most recent call last):
  File "c:/docume~1/ashutosh/locals~1/temp/tmpvcvr4w.nl", line 260, in ?
    runtime_parameters = runtime_parameters
ATKError: Exceeded maximum number of self-consistent iterations.
Terminated Abnormally

Please tell me what does this error asks me to refine?

Thanks.

9

General Questions and Answers / Re: From crystalmaker to VNL

« on: July 31, 2009, 20:13 »

That's a good idea.

10

General Questions and Answers / Re: From crystalmaker to VNL

« on: July 31, 2009, 11:22 »

I haven't tried that but I think it's simple. After the 40th-50th iteration (as in my case), if I see the properties using .nc file if they are anywhere close to the result for a standard-solved problem then my parameters would be close to correct. Then a few changes I would do accordingly when I run my actual structure. I might be wrong in this approach and so you are experienced to tell me if I actually am!

Thanks.

11

General Questions and Answers / Re: Regarding Two-probe convergence

« on: July 31, 2009, 11:16 »

Thanks Anders.
Yeah, in previous calculations it went upto 140 iterations and days to converge.

I stopped this previous simulation as I felt it is going nowhere. Will run it on the cluster then to see if it converges quickly.

12

General Questions and Answers / Re: From crystalmaker to VNL

« on: July 31, 2009, 10:06 »

Yeah, I meant to check the properties in the result browser after stopping the iterations at an intermediate stage.  I would do this for a standard system before trying it on my structure.

13

General Questions and Answers / Re: Regarding Two-probe convergence

« on: July 31, 2009, 10:03 »

Thanks Anders.
The calculations are still running, so no output file. I tried attaching the .nc file that has still not converged. But it exceeds the allowed attachment limit of 512KB as it popped-up. The count is 35 iterations (it is running at the rate of approx. 4 iterations per hour!!).

The dRho is still fluctuating from E-1 to E-3 from 20th iteration upto 35th. and same goes for the dEtot values. Do you think these are not converging?? I guess so.

I think I would switch over to a cluster. One quick question, since the cluster that I have access to is Linux base, so will this work if I prepare the script file on VNL and run that .py file on ATK? Just to save myself from scripting???

Thankyou.

14

General Questions and Answers / Regarding Two-probe convergence

« on: July 31, 2009, 02:35 »

Can someone please take a look into this file and let me know what improvement I can make in the initial-parameters for a better convergence(and probably in shorter time). Right it has been 40 hours and the calculations are still running  

My system: Processor: AMD Athlon(tm) 64 X2 Dual Core 4400+
                 RAM: 3GB

Thankyou.

15

General Questions and Answers / Re: From crystalmaker to VNL

« on: July 31, 2009, 01:34 »

I got it done. Thanks, Anders.

The other question I have is: After we have started the SCF calculations on VNL, and take the following steps:

1st, We stop the calculations in between(before it gets terminated normally itself)
2nd, Restore the check-point file
3rd, Do some properety-calculations(like transmission spectrum, DOS, etc.)
4th, Analyze from the results where to the calculations are leading.
5th, If good, can restore the .nc file to carry-on the SCF calculations from where we stopped in the 1st step.

Will this in anyway deviate the results?

This in order to ensure that we don't wait long enough for our calculations that might terminate abnormally/ or to analyze the effect of various initializing parameters in our structures.