Reproduction of YAMBO Paper Results

[c3031929]

Hi,

I've been trying to reproduce the results for bulk Si in the paper Marini, A., Hogan, C., Gruning, M., Varsano, D., "yambo: An ab initio tool for excited state calculations", Computer Physics Communications, 180 (2009) 1392. Unfortunately I have as yet been unable to reproduce the results. On my system YAMBO is compiled without NETCDF support and as such I can't even check my calculations against the pre-computed structure supplied in the yambo compilation (since it is in NETCDF form only).

My aim has been this: to compute the bulk Si results found in the paper starting from the beginning at Abinit. I have tried almost every possible thing I can think of and still I am unable to reproduce the results published i the paper. In my attempts to do this I have examined the ecut value, the number of bands, the number of k-points and the lattice constant. None of these were significant enough to bring my outputs into line with the papers results. The only thing I did do that helped in any significant way was to change the way I defined the unit cell so that the rprim components were all positive for both the atoms (instead of having the second atom with negative x and y components). (Any hints as to why that made a significant difference would also be appreciated).

I have attached the Abinit files that I start from as well as all the files associated with the Yambo run. I based my Abinit file on the input Abinit file for the Solid_LiF tutorial with appropriate modifications.

Any help here would be most appreciated.

[myrta gruning]

Hallo Daniel

Can you please edit the signature in your user profile adding your name and affiliation (so they will appear automatically at the bottom of all your posts)? I had to search the forum history for your name...

There are some differences in your abinit input: the cutoff (for the sample in the source is 15Ry), and the geometry, acell is 10.183 au
About changing the rprim causing changes in your output, I guess it corresponds to rotate the sample, so finally equivalent to changing the direction of the electric field.
Another think that I noticed, that may produce some difference is that in the sample there is no (n)s.kb_pp. So to reproduce the results you can try to remove it from the SAVE dir.
I hope this will help you. Let us know.

Final advice : try to compile with netcdf

Regards,
Myrta

[c3031929]

Myrta,

Sorry for the slow response. It took me a while to get Yambo successfully installed with NetCDF. When I did the calculations again, there was no change with NetCDF as there was without. I did find the error though. Apparently I had been entering one number incorrectly. I was entering 0.2 instead of 0.02 into the BDmRange variable.

Thanks for the help.

Cheers,

Danno