Want to rescaling vectors in converter

JaeHyung Yu

unread,

Mar 30, 2016, 1:27:37 PM3/30/16

to QWalk user's group

Hi,

I want to convert DFT result of my system, which has lattice constant of 6 angstrom x 6 angstrom x 100 angstrom.

Because I am doing some layered structure calculation, there are vacuum space in c-direction.

However, I had a message from crystal2qmc converter :

"WARNING: rescaling lattice vector to 100 from 188.973. This will probably not change your results, but if it does, change EWALD_GMAX in the system input to something higher."

It seems that converter rescaled the vector in c direction (which has 100 angstrom).

When I tried to see atomic orbitals by PLOT command, rescaled geometries confuses me.

Can I avoid rescaling during converting my DFT output?

Thank you!

Best,

Jaehyung Yu

Lucas Wagner

unread,

Mar 30, 2016, 3:15:39 PM3/30/16

to qwalk...@googlegroups.com

Hi Jae,

Why do you say that rescaled geometries confuse you? I don't understand what that means.

It changes the lattice constant so that the 3D Ewald summation converges faster, but that shouldn't change anything about the calculation.

Best,

Lucas

--
You received this message because you are subscribed to the Google Groups "QWalk user's group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qwalk-users...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

JaeHyung Yu

unread,

Mar 30, 2016, 4:50:59 PM3/30/16

to QWalk user's group

Hi Lucas,

I know it will not influence to the calculation.

However, I want to see atomic orbitals without rescaling, in order to choose orbitals where excited electron will be.

But when converter rescales lattice vectors, atomic structures seems overlapped when I use PLOT command.

Because of that, I cannot distinguish orbitals - that's why I just wonder how to avoid rescaling.

Thank you.

2016년 3월 30일 수요일 오후 2시 15분 39초 UTC-5, Lucas 님의 말:

Lucas Wagner

unread,

Mar 30, 2016, 5:06:42 PM3/30/16

to qwalk...@googlegroups.com

Jae, that shouldn't happen. Could you post your files?

Lucas

JaeHyung Yu

unread,

Mar 31, 2016, 6:27:01 PM3/31/16

to QWalk user's group

Hi Lucas,

1.

Due to file size, I posted my input file of CRYSTAL14(name : Crystal_Input_MoS2) , and also converted qwalk input ( you can download from here : https://drive.google.com/file/d/0B6eLwh9HuZvUc2h0OUd6RmFscTg/view?usp=sharing ).

In CRYSTAL file, two MoS2 layers are separated by ~50 A.

this is crystal geometry that i used, in cartesian coordinate with unit of A(angstrom)

242	MO	-0.913354486	-1.581987008	3.075511422
242	MO	-0.913354486	-1.581987008	-46.92448858
242	MO	-0.913354486	1.582012992	3.075511422
242	MO	-0.913354486	1.582012992	-46.92448858
242	MO	1.826749891	-3.163987008	3.075511422
242	MO	1.826749891	-3.163987008	-46.92448858
242	MO	1.826749891	1.29919E-05	3.075511422
242	MO	1.826749891	1.29919E-05	-46.92448858
216	S	0.913361242	1.581993446	1.626382349
216	S	0.913361242	1.581993446	-48.37361765
216	S	0.913361242	-1.582006554	1.626382349
216	S	0.913361242	-1.582006554	-48.37361765
216	S	-1.826743135	3.163993446	1.626382349
216	S	-1.826743135	3.163993446	-48.37361765
216	S	-1.826743135	-6.55355E-06	1.626382349
216	S	-1.826743135	-6.55356E-06	-48.37361765
216	S	0.913361314	1.581993373	4.524606079
216	S	0.913361314	1.581993373	-45.47539392
216	S	0.913361314	-1.582006627	4.524606079
216	S	0.913361314	-1.582006627	-45.47539392
216	S	-1.826743064	3.163993373	4.524606079
216	S	-1.826743064	3.163993373	-45.47539392
216	S	-1.826743064	-6.62663E-06	4.524606079
216	S	-1.826743064	-6.62663E-06	-45.47539392

However, the rescaling procedure of converter reduces separation between two layer, to almost overlap each other.

This is geometry shown in converted qwalk input file.

ATOM { MO 6 COOR 9.924615571 3.736957995 18.31187431 }

ATOM { MO 6 COOR 9.924615571 3.736957995 23.82556761 }

ATOM { MO 6 COOR 9.924615571 -2.242135517 18.31187431 }

ATOM { MO 6 COOR 9.924615571 -2.242135517 23.82556761 }

ATOM { MO 6 COOR 4.746568722 6.726504751 18.31187431 }

ATOM { MO 6 COOR 4.746568722 6.726504751 23.82556761 }

ATOM { MO 6 COOR 4.746568722 0.7474112387 18.31187431 }

ATOM { MO 6 COOR 4.746568722 0.7474112387 23.82556761 }

ATOM { S 6 COOR 3.020514322 3.736921047 15.57341723 }

ATOM { S 6 COOR 3.020514322 3.736921047 21.08711053 }

ATOM { S 6 COOR 3.020514322 9.716014559 15.57341723 }

ATOM { S 6 COOR 3.020514322 9.716014559 21.08711053 }

ATOM { S 6 COOR 8.19856117 0.7473742908 15.57341723 }

ATOM { S 6 COOR 8.19856117 0.7473742908 21.08711053 }

ATOM { S 6 COOR 8.19856117 6.726467803 15.57341723 }

ATOM { S 6 COOR 8.19856117 6.726467803 21.08711053 }

ATOM { S 6 COOR 3.020514457 3.736920909 21.05026635 }

ATOM { S 6 COOR 3.020514457 3.736920909 26.56395966 }

ATOM { S 6 COOR 3.020514457 9.716014421 21.05026635 }

ATOM { S 6 COOR 3.020514457 9.716014421 26.56395966 }

ATOM { S 6 COOR 8.198561305 0.7473741527 21.05026635 }

ATOM { S 6 COOR 8.198561305 0.7473741527 26.56395966 }

ATOM { S 6 COOR 8.198561305 6.726467665 21.05026635 }

ATOM { S 6 COOR 8.198561305 6.726467665 26.56395966 }

As you can see above, two S atoms becomes almost overlapped - it also can be seen at atomic orbital plot.

(I want to upload all my input files, but I cannot because of file size limits. I am sorry for that.)

I used qwalk parallel version, which is downloaded from Github at January 30th in this year.

2. (Related input files cam be downloaded from here : https://drive.google.com/file/d/0B6eLwh9HuZvUc2h0OUd6RmFscTg/view?usp=sharing )

Also, during my other calculation, I have errors in DMC calculation.

This calculation is almost same calculation with above, but I reduced vacuum space to avoid rescaling.

I successfully avoid rescaling, and optimize jastrow factor also successful.

So I tried DMC calculation my input file, but DMC calculation has error :

(used parallel version Qwalk in Blue Water machine, with 2048 cores, with same version of above.)

cutoff 13.3473

total centers 294

cutoff 13.3473

total centers 294

Ewald sum using 4482 reciprocal points

Error Did over 1000 shifts and we're still out of the simulation cell. There's probably something wrong. Position : nan

Firstly I thought it is because of vacuum space in my calculation system, but other system which also contain vacuum space worked well.

And also I searched in our user group, there was 1 posting (title: Questions - version 0.96.0) on same error message; but that was VMC calculation on the personal computer, and the source of error is bug in the code.

So, How can I avoid this position error?

Thank you!

Best,

Jaehyung Yu

2016년 3월 30일 수요일 오후 4시 6분 42초 UTC-5, Lucas 님의 말:

Crystal_Input_MoS2

JaeHyung Yu

unread,

Mar 31, 2016, 6:28:38 PM3/31/16

to QWalk user's group

Sorry for inconvenience,

file related to 2nd problem can be downloaded from here :

https://drive.google.com/file/d/0B6eLwh9HuZvUNVNNY3d3QjhPcWc/view?usp=sharing

2016년 3월 31일 목요일 오후 5시 27분 1초 UTC-5, JaeHyung Yu 님의 말:

Reply all

Reply to author

Forward