[Pw_forum] Problems with interrupted electron-phonon coupling calculation
I just thought I will restate my question, in case it wasn't clear last time. In the case of a phonon only calculation at a particular q point, it is possible to run the phonon calculation for different irreps separately and then copy the data-file.q.irr.xml files into a common place and then run the phonon calculation again after setting recover = .true., to obtain the dynamical matrix for that q point. This is essentially grid parallelization. I need do this for an electron-phonon calculation. Could someone help me with what output files that I have to copy from the runs on different irreps so that I can get the correct electron-phonon coupling results when I run all irreps together after setting recover = .true.? Thank you and I would really appreciate any help. Sridhar Graduate Student Purdue University On Fri, Mar 7, 2014 at 1:45 PM, Sridhar Sadasivam wrote: > I am using espresso 5.0.3. Thanks, > Sridhar > > > On Fri, Mar 7, 2014 at 1:30 PM, Paolo Giannozzi uniud.it>wrote: > >> Code version? P. >> >> On Thu, 2014-03-06 at 23:36 +0530, Sridhar Sadasivam wrote: >> > Hello, >> > >> > >> > I am working on an electron-phonon coupling calculation of a large >> > system (30 atoms). At each q-point I will have 30 x 3 = 90 irreducible >> > representations. This will take a lot of computer time and I expect >> > that my calculation will be interrupted (due to max wall time >> > limitations in the computing cluster I use). So I will need to restart >> > the calculation many times. >> > >> > >> > To test if everything runs fine, I tried to run an electron-phonon >> > coupling calculation on a simple 1 atom system. For a particular >> > q-point I had 2 irreducible representations and I interrupted my >> > calculation after the self-consistent calculation for the 1st >> > representation. I then restarted the calculation setting "recover >> > = .true.". The calculation starts with the 2nd representation and >> > seems to complete, but the electron-phonon linewidths are wrong, i.e., >> > the results don't match the results I get when I run without any >> > interruption. The phonon frequencies are correct in the interrupted >> > calculation but the electron-phonon results are not. For some reason, >> > restarting an interrupted ph.x calculation does not give the correct >> > results for electron-phonon coupling. >> > >> > >> > Does anyone know how I can resolve this issue? This is crucial as I >> > will not be able to run the large system (30 atoms) without having to >> > restart many times. >> > >> > >> > I would really appreciate if someone could help me out on this. Please >> > let me know if my question isn't clear and I can attach output files >> > as well. >> > Regards, >> > Sridhar >> > Purdue University, >> > USA >> > >> > >> > >> > >> > >> > >> > >> > >> > ___ >> > Pw_forum mailing list >> > Pw_forum at pwscf.org >> > http://pwscf.org/mailman/listinfo/pw_forum >> >> -- >> Paolo Giannozzi, Dept. Chemistry&Physics&Environment, >> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy >> Phone +39-0432-558216, fax +39-0432-558222 >> >> ___ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://pwscf.org/mailman/listinfo/pw_forum >> > > -- next part -- An HTML attachment was scrubbed... URL: http://pwscf.org/pipermail/pw_forum/attachments/20140309/63f72d17/attachment.html
[Pw_forum] Problems with interrupted electron-phonon coupling calculation
I am using espresso 5.0.3. Thanks, Sridhar On Fri, Mar 7, 2014 at 1:30 PM, Paolo Giannozzi wrote: > Code version? P. > > On Thu, 2014-03-06 at 23:36 +0530, Sridhar Sadasivam wrote: > > Hello, > > > > > > I am working on an electron-phonon coupling calculation of a large > > system (30 atoms). At each q-point I will have 30 x 3 = 90 irreducible > > representations. This will take a lot of computer time and I expect > > that my calculation will be interrupted (due to max wall time > > limitations in the computing cluster I use). So I will need to restart > > the calculation many times. > > > > > > To test if everything runs fine, I tried to run an electron-phonon > > coupling calculation on a simple 1 atom system. For a particular > > q-point I had 2 irreducible representations and I interrupted my > > calculation after the self-consistent calculation for the 1st > > representation. I then restarted the calculation setting "recover > > = .true.". The calculation starts with the 2nd representation and > > seems to complete, but the electron-phonon linewidths are wrong, i.e., > > the results don't match the results I get when I run without any > > interruption. The phonon frequencies are correct in the interrupted > > calculation but the electron-phonon results are not. For some reason, > > restarting an interrupted ph.x calculation does not give the correct > > results for electron-phonon coupling. > > > > > > Does anyone know how I can resolve this issue? This is crucial as I > > will not be able to run the large system (30 atoms) without having to > > restart many times. > > > > > > I would really appreciate if someone could help me out on this. Please > > let me know if my question isn't clear and I can attach output files > > as well. > > Regards, > > Sridhar > > Purdue University, > > USA > > > > > > > > > > > > > > > > > > ___ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://pwscf.org/mailman/listinfo/pw_forum > > -- > Paolo Giannozzi, Dept. Chemistry&Physics&Environment, > Univ. Udine, via delle Scienze 208, 33100 Udine, Italy > Phone +39-0432-558216, fax +39-0432-558222 > > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://pwscf.org/mailman/listinfo/pw_forum > -- next part -- An HTML attachment was scrubbed... URL: http://pwscf.org/pipermail/pw_forum/attachments/20140307/8f33a25d/attachment.html
[Pw_forum] Problems with interrupted electron-phonon coupling calculation
Code version? P. On Thu, 2014-03-06 at 23:36 +0530, Sridhar Sadasivam wrote: > Hello, > > > I am working on an electron-phonon coupling calculation of a large > system (30 atoms). At each q-point I will have 30 x 3 = 90 irreducible > representations. This will take a lot of computer time and I expect > that my calculation will be interrupted (due to max wall time > limitations in the computing cluster I use). So I will need to restart > the calculation many times. > > > To test if everything runs fine, I tried to run an electron-phonon > coupling calculation on a simple 1 atom system. For a particular > q-point I had 2 irreducible representations and I interrupted my > calculation after the self-consistent calculation for the 1st > representation. I then restarted the calculation setting "recover > = .true.". The calculation starts with the 2nd representation and > seems to complete, but the electron-phonon linewidths are wrong, i.e., > the results don't match the results I get when I run without any > interruption. The phonon frequencies are correct in the interrupted > calculation but the electron-phonon results are not. For some reason, > restarting an interrupted ph.x calculation does not give the correct > results for electron-phonon coupling. > > > Does anyone know how I can resolve this issue? This is crucial as I > will not be able to run the large system (30 atoms) without having to > restart many times. > > > I would really appreciate if someone could help me out on this. Please > let me know if my question isn't clear and I can attach output files > as well. > Regards, > Sridhar > Purdue University, > USA > > > > > > > > > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://pwscf.org/mailman/listinfo/pw_forum -- Paolo Giannozzi, Dept. Chemistry&Physics&Environment, Univ. Udine, via delle Scienze 208, 33100 Udine, Italy Phone +39-0432-558216, fax +39-0432-558222
[Pw_forum] Problems with interrupted electron-phonon coupling calculation
Hello, I am working on an electron-phonon coupling calculation of a large system (30 atoms). At each q-point I will have 30 x 3 = 90 irreducible representations. This will take a lot of computer time and I expect that my calculation will be interrupted (due to max wall time limitations in the computing cluster I use). So I will need to restart the calculation many times. To test if everything runs fine, I tried to run an electron-phonon coupling calculation on a simple 1 atom system. For a particular q-point I had 2 irreducible representations and I interrupted my calculation after the self-consistent calculation for the 1st representation. I then restarted the calculation setting "recover = .true.". The calculation starts with the 2nd representation and seems to complete, but the electron-phonon linewidths are wrong, i.e., the results don't match the results I get when I run without any interruption. The phonon frequencies are correct in the interrupted calculation but the electron-phonon results are not. For some reason, restarting an interrupted ph.x calculation does not give the correct results for electron-phonon coupling. Does anyone know how I can resolve this issue? This is crucial as I will not be able to run the large system (30 atoms) without having to restart many times. I would really appreciate if someone could help me out on this. Please let me know if my question isn't clear and I can attach output files as well. Regards, Sridhar Purdue University, USA -- next part -- An HTML attachment was scrubbed... URL: http://pwscf.org/pipermail/pw_forum/attachments/20140306/4aa329e1/attachment.html
