G1-G2 pairs outside the G-sphere in screening calculation  [SOLVED]

GW, Bethe-Salpeter …

Moderators: maryam.azizi, bruneval

Locked
wxiaom86
Posts: 14
Joined: Tue Nov 29, 2016 6:28 pm

G1-G2 pairs outside the G-sphere in screening calculation

Post by wxiaom86 » Sat May 27, 2017 11:01 pm

Hello,

Does anyone know how to solve the below problem during screening calculation using the gwcomp technique?
''gwcompleteness.F90:149:WARNING
Number of G1-G2 pairs outside the G-sphere for Wfns = 396572"

It is suggested to use either larger ecutwfn or smaller ecuteps to get rid of that. However, my ecuteps is 5 Ry already small and ecutwfn is the default of ecut which cannot be larger. Any idea?

Best,
xW

wxiaom86
Posts: 14
Joined: Tue Nov 29, 2016 6:28 pm

Re: G1-G2 pairs outside the G-sphere in screening calculatio

Post by wxiaom86 » Sun May 28, 2017 6:44 am

I also tried to increase ecutwfn to 40 and 50 Ha, the number of G1-G2 pairs did decrease, but still a huge number, about 1e+5. Anything else I can try?

wxiaom86
Posts: 14
Joined: Tue Nov 29, 2016 6:28 pm

Re: G1-G2 pairs outside the G-sphere in screening calculatio  [SOLVED]

Post by wxiaom86 » Fri Dec 08, 2017 5:42 pm

This problem may be due to the large anisotropy of my system. I have a very long crystal length.
wxiaom86 wrote:Hello,

Does anyone know how to solve the below problem during screening calculation using the gwcomp technique?
''gwcompleteness.F90:149:WARNING
Number of G1-G2 pairs outside the G-sphere for Wfns = 396572"

It is suggested to use either larger ecutwfn or smaller ecuteps to get rid of that. However, my ecuteps is 5 Ry already small and ecutwfn is the default of ecut which cannot be larger. Any idea?

Best,
xW

yow5110
Posts: 3
Joined: Wed Mar 20, 2019 4:42 pm

Re: G1-G2 pairs outside the G-sphere in screening calculatio

Post by yow5110 » Wed Mar 20, 2019 4:59 pm

1 year too late maybe:

Getting G1-G2 vectors outside the G-sphere for wfn is fine I think, but the question is whether the summation in the gwcomp method is converged (when summing over G vectors inside the G-sphere). I just made a post that might affect the results that you're getting.
viewtopic.php?f=11&t=4071

Best
Yuanxi Wang

User avatar
hameed1987
Posts: 41
Joined: Sat Jun 22, 2019 12:45 pm
Location: The UK

Re: G1-G2 pairs outside the G-sphere in screening calculation

Post by hameed1987 » Mon Mar 23, 2020 6:57 pm

Dear wxiaom86
My problem is similar to yours. Please tell me how did you solve it ? please see below my error

--- !ERROR
src_file: kgindex.F90
src_line: 110
mpi_rank: 2
message: |
The G-vector with ig: 4436 falls outside the FFT box.
Dear All,


WR
Hameed

Locked