Dear Iurii and Dario, Thanks for your answers. At some point I gave up with the Davidson code since it was impossible to converge even just one excited state, this is why I switched to Lanczos. I had some hope on being able to understand the peaks but now I see that I've to try again with Davidson.
All my best, Sergi 2016-05-20 11:07 GMT+02:00 dario rocca <[email protected]>: > Dear Sergi, > I would like just to add that in the examples folder of the turboTDDFT > code you can find examples on > how to perform the Davidson diagonalization. > Best, > Dario Rocca > > On Fri, May 20, 2016 at 11:00 AM, Timrov Iurii <[email protected]> > wrote: > >> Dear Dr. Sergi Vela, >> >> >> With the current implementation of the Lanczos algorithm in the >> turboTDDFT program it is not possible to perform analysis of the origin of >> individual peaks in the absorption spectrum. However, such an analysis can >> be done using another flavor of the turboTDDFT code, namely the >> turbo_davidson.x program. More information can be found in this >> publication: >> >> >> X. Ge, S. J. Binnie, D. Rocca, R. Gebauer, and S. Baroni, "turboTDDFT >> 2.0—Hybrid functionals and new algorithms within time-dependent >> density-functional perturbation theory", Comput. Phys. Commun. 185, 2080 >> (2014). >> >> >> With the so-called turboDavidson code (turbo_davidson.x) you can >> compute the absorption spectrum (the same as with Lanczos) but as a >> plus in the output file you will find the information about the origin of >> the peaks (from which occupied level to which empty level there was a >> transition and what is the corresponding intensity etc.). However, >> if you are interested in computing the spectrum in a very >> wide energy range, then turboDavidson becomes too expensive, and hence it >> is recommended to use the Lanczos algorithm (at the cost described above, >> i.e. no information about the origin of the peaks). >> >> >> HTH >> >> >> Best regards, >> >> Iurii >> >> >> -- >> Dr. Iurii Timrov >> Postdoctoral Researcher >> Swiss Federal Institute of Technology Lausanne (EPFL) >> Laboratory of Theory and Simulation of Materials (THEOS) >> CH-1015 Lausanne, Switzerland >> +41 21 69 34 881 >> http://people.epfl.ch/265334 >> ------------------------------ >> *From:* [email protected] <[email protected]> on >> behalf of Sergi Vela <[email protected]> >> *Sent:* Friday, May 20, 2016 10:30 AM >> *To:* PWSCF Forum >> *Subject:* [Pw_forum] analysis of TDDFPT results >> >> Dear QE users and developers, >> >> I've successfully performed a TDDFPT (Lanczos approach) calculation on a >> system of my interest, and now I would like to analyze the different bands >> of the resulting spectrum. I would like to ascribe them to chemical >> (orbital-excitations, charge-transfer) processes such as I would do when >> using standard TDDFT codes. >> >> I'd like to know if this is indeed possible and, if so, I'd appreciate >> some information on how to do it. >> >> All my best, >> Dr. Sergi Vela, >> Laboratoire de Chimie Quantique, Universitè de Strasbourg, France >> >> _______________________________________________ >> Pw_forum mailing list >> [email protected] >> http://pwscf.org/mailman/listinfo/pw_forum >> > > > _______________________________________________ > Pw_forum mailing list > [email protected] > http://pwscf.org/mailman/listinfo/pw_forum >
_______________________________________________ Pw_forum mailing list [email protected] http://pwscf.org/mailman/listinfo/pw_forum
