GW CORRECTION.

[fatimazahra]

Hello
1) I'm a PHd student, I just used yambo code for GW with PWscf code but I have a problem: my programe does not give me the output file, it gives me juste the file: r_HF_and_locXC_em1d_ppa_gw0

r_HF_and_locXC_em1d_ppa_gw0.txt

, but my programe doesn't give me the o.-GoWo_PPA.qp file.
I have done the following steps:
Firstly, I converted with p2y
then I ran the programme with yambo -g n -p p
2) in r_HF_and_locXC_em1d_ppa_gw0 ,
Indirect Gaps [ev]: 2.261132 4.610378
[spin up]: 2.261279 4.610509
[spin dn]: 2.261132 4.610378
Direct Gaps [ev]: 2.261132 6.305942
[spin up]: 2.261279 6.305998
[spin dn]: 2.261132 6.305942

which of these values corresponds to the approximation GGA or GW ?

[Daniele Varsano]

Dear Fatima,
first let me ask to fill your signature with your affiliation, this is a rule of the forum.
From your report I can see that your calculation did not finished. Did you receive any error message in the standard output?
have a look to the l_HF_.., or error messages produced by the queue system id you are running in a cluster.
The gap you found in the report, is the energy gap or your previous DFT calculation read from Pwscf, I presume GGA.
Next, I can see that you included just one empty state, in the calculations needed for GW (calculation of the screening and GW summation) you have a sum over state, so many empty state have to be included.

Best,
Daniele

[fatimazahra]

I left the program 4 days but doesn't want to finish.
in scf and nscf i don't add the force_symmorphic=.true., I read in forum that I should add in the input, is it possible that this is the problem ?

[Daniele Varsano]

Dear Fatima,

Code: Select all

force_symmorphic=.true

is a good practice to include but I do not think this is the problem, as in this case the code should stop immediately. In order to understand what is happening, please post the r_HF_and_locXC_em1d_ppa_gw0 file as well as your input file.

Best,

Daniele

[fatimazahra]

Hello
my input file

yambo.in.txt

and output

r_HF_and_locXC_em1d_ppa_gw0_05.txt

Thank you very much

[Daniele Varsano]

Dear Fatima,
sorry for the mistype, in order ro understand what is going on, I was asking the standard output and not the report:
l_HF_and_locXC_em1d_ppa_gw0** instead of r_HF_and_locXC_em1d_ppa_gw0_05.txt.

Anyway, even if not related to the problem of not-ending calculation:
please note: you calculated only 12 states, ie only one conduction band, that surely is not enough to get converged results.
You are including only one Gvector: NGsBlkXp= 1 with is surely not enough
All this variables should be tested until convergence.

Best,
Daniele

[fatimazahra]

i.e should be changed these variables randomly ?
Thank you for all

[Daniele Varsano]

Dear Fatima,

i.e should be changed these variables randomly ?
Thank you for all

Well, not randomly. You should increase the values of the variables until the results does not change anymore (according to the precision you want achieve). Of course, the bigger the values, the heavier is the calculation in terms of time and memory. Anyway the problem that the run does not end, it is not related to that. In order to understand what is happening post the l_HF_and_locXC_em1d_ppa_gw0* file.

Best,
Daniele

[fatimazahra]

There isn’t a file named l_HF_and_locXC_em1d_ppa_gw0, there is just r_HF_and_locXC_em1d_ppa_gw0
Anyway I will run the program with these changes,
Thank you very match for all

[Daniele Varsano]

Dear Fatima,
that sounds very strange!! the l_HF_xx file should be somewhere. Anyway, run the code again and if it fails, post here all the information.

Good luck,
Daniele