Re: [Wien] Error in Hub-U calculation
I don't know for sure, but it might be that the /localscratch permission denied errors make the parallel calculations broken. To fix that, the cluster administer would probably have to give you file permission to use /localscratch or /localscratch would have to be changed to a directory that you have access to. I believe the /localscratch can be changed by running ./siteconfig and selecting "T Temp Path" in the menu. However, it seems that the cluster administrator would have to do that. Since, if you were able to do that, it seems potentially possible for you to upgrade/install the latest WIEN2k version (18.2) to have the +U (orb) serial to parallel fix [ https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg17587.html ]. While waiting more than two months for a cluster administrator as you say, it maybe that you can only do serial calculations on the cluster. If your still using the job script (job1) in the post https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg17586.html that appears to be trying to use mpi for lapw0 and k-point parallel for lapw1/2. The dayfile in that post has "running lapw0 in single mode" indicated that mpi lapw0 is not available or not working such that it reverted to using its serial calculation mode. That same dayfile appears to be from perhaps multiple continuation attempts with the -NI switch (or a save_lapw between "runsp_lapw -p" and "runsp_lapw -p -orb") as I see only "lapw0 -p" followed by "orb -up -p". Thus, it seems to have lost possibly helpful and important previous scf cycle output that came before it. For example, for a single mode lapw0 and k-point parallel lapw1/2 using only the two commands [1]: init_lapw -b -sp (with only case.struct, case.indm, and case.inorb files in the case folder) runsp_lapw -p -orb A typical case.dayfile for that shows an entire scf cycle (lapw0, lapw1, lapw2, etc.) before it does "orb -up -p": Calculating Utest in /home/username/wiendata/Utest on computername with PID 19009 using WIEN2k_17.1 (Release 30/6/2017) in /home/username/WIEN2k start (Tue Aug 7 22:52:41 MDT 2018) with lapw0 (40/99 to go) cycle 1 (Tue Aug 7 22:52:41 MDT 2018) (40/99 to go) > lapw0 -p (22:52:41) starting parallel lapw0 at Tue Aug 7 22:52:41 MDT 2018 .machine0 : processors running lapw0 in single mode 40.9u 0.2s 0:41.77 98.7% 0+0k 7728+2760io 25pf+0w > lapw1 -up -p -orb (22:53:23) starting parallel lapw1 at Tue Aug 7 22:53:23 MDT 2018 -> starting parallel LAPW1 jobs at Tue Aug 7 22:53:23 MDT 2018 running LAPW1 in parallel mode (using .machines) 2 number_of_parallel_jobs localhost(24) 233.0u 3.7s 3:58.10 99.4% 0+0k 9792+82832io 40pf+0w localhost(23) 220.7u 3.5s 3:45.06 99.6% 0+0k 0+77400io 0pf+0w Summary of lapw1para: localhost k=47 user=453.7 wallclock=463.16 454.2u 7.4s 3:59.43 192.8% 0+0k 10048+160632io 43pf+0w > lapw1 -dn -p -orb (22:57:22) starting parallel lapw1 at Tue Aug 7 22:57:22 MDT 2018 -> starting parallel LAPW1 jobs at Tue Aug 7 22:57:22 MDT 2018 running LAPW1 in parallel mode (using .machines.help) 2 number_of_parallel_jobs localhost(24) 233.3u 3.4s 3:57.62 99.6% 0+0k 0+83072io 0pf+0w localhost(23) 221.3u 3.3s 3:45.34 99.6% 0+0k 0+77528io 0pf+0w Summary of lapw1para: localhost k=47 user=454.6 wallclock=462.96 455.0u 6.9s 3:58.84 193.4% 0+0k 0+160976io 0pf+0w > lapw2 -up -p -orb (23:01:21) running LAPW2 in parallel mode localhost 69.5u 0.8s 1:10.72 99.5% 0+0k 0+1408io 0pf+0w localhost 69.1u 0.7s 1:10.08 99.6% 0+0k 1376+1312io 4pf+0w Summary of lapw2para: localhost user=138.6 wallclock=140.8 139.5u 1.6s 1:12.99 193.4% 0+0k 5952+6176io 19pf+0w > lapw2 -dn -p -orb (23:02:34) running LAPW2 in parallel mode localhost 59.4u 0.8s 1:00.56 99.6% 0+0k 0+1408io 0pf+0w localhost 59.6u 0.7s 1:00.54 99.7% 0+0k 0+1312io 0pf+0w Summary of lapw2para: localhost user=119 wallclock=121.1 120.0u 1.6s 1:02.75 193.8% 0+0k 0+6160io 0pf+0w > lcore -up (23:03:37) 0.0u 0.0s 0:00.13 61.5% 0+0k 1688+552io 6pf+0w > lcore -dn (23:03:37) 0.0u 0.0s 0:00.08 100.0% 0+0k 0+552io 0pf+0w > mixer -orb (23:03:38) 0.3u 0.0s 0:00.51 70.5% 0+0k 4864+3680io 18pf+0w :ENERGY convergence: 0 0.0001 0 :CHARGE convergence: 0 0. 0 cycle 2 (Tue Aug 7 23:03:38 MDT 2018) (39/98 to go) > lapw0 -p (23:03:38) starting parallel lapw0 at Tue Aug 7 23:03:38 MDT 2018 .machine0 : processors running lapw0 in single mode 45.2u 0.1s 0:45.54 99.7% 0+0k 8+2760io 0pf+0w > orb -up -p (23:04:24) 0.0u 0.0s 0:00.05 0.0% 0+0k 1464+32io 5pf+0w > orb -dn -p (23:04:24) 0.0u 0.0s 0:00.01 0.0% 0+0k 0+32io 0pf+0w > lapw1 -up -p -orb (23:04:24) starting parallel lapw1 at Tue Aug 7 23:04:24 MDT 2018 -> starting parallel LAPW1 jobs at Tue Aug 7 23:04:24 MDT 2018 running LAPW1 in parallel mode
Re: [Wien] Error in Hub-U calculation
Did you already try to run an L(S)DA+U calculation without restarting from a previous L(S)DA one ? Also, I am not sure the -dm option is necessary in this case. W. De: "shaymlal dayananda" À: "A. Mailing List for WIEN2k Users" Envoyé: Mardi 7 Août 2018 14:49:33 Objet: [Wien] Error in Hub-U calculation On Tuesday, August 7, 2018 12:41 PM, shaymlal dayananda wrote: Dear Developers and users I am in a trouble that I could not recover my problem yet. I tried all of your suggestions. I am summarizing them below. 1. I finished runsp_lapw -NI -p -ec 0.0001 correctly and saved the files. 2. I added case.indmc and case.inorb files ( I have copied them one of previous emails) 3. submitted the job runsp_lapw -NI -p -dm -orb -ec 0.0001 4. This job stopped with the same previous errors. Anyway I am coping them again along with case.vorbdef STDOUT cp: cannot create regular file '/localscratch//.tmp_lapw1para.30532': Permission denied cp: cannot stat '/localscratch//.tmp_lapw1para.30532': No such file or directory /localscratch//.tmp_testpara_new.30532_2: Permission denied. grep: /localscratch//.tmp_lapw1para.30532: No such file or directory cut: /localscratch//.tmp_testpara_new.30532_2: No such file or directory forrtl: severe (24): end-of-file during read, unit 7, file /scratch/WIEN2k17/TEST/T-hubU/T-hubU.vorbup Image PC Routine Line Source lapw1c 00461858 Unknown Unknown Unknown lapw1c 00497DDA Unknown Unknown Unknown lapw1c 0042C543 inilpw_ 276 inilpw.f lapw1c 0042F302 MAIN__ 42 lapw1_tmp_.F lapw1c 00403FEE Unknown Unknown Unknown libc.so.6 2AD5081B12E0 Unknown Unknown Unknown lapw1c 00403EEA Unknown Unknown Unknown 5. next I tried: energyup/dn copied to energyup_1/dn_1, all dmat* removed and then tried runsp_lapw -NI -p -dm -orb -ec 0.0001 But it gave me the same error. I has changed the energyup_1/dn_1 (empty now) 6. I tried running x lapw1 with deleting dmat* as William as sugested, but this also leads for the same error as I have given in 5. above. I am very much appreciate if anyone give the correct way to do hubbard U included PARALLEL calculation with Wien2k 17.1. I have asked our supercomputers to install the latest version. But due to internal issue this will be late more than two months more. I cannot wait that long. Thank you Daya On Tuesday, June 26, 2018 2:52 PM, William Lafargue-dit-Hauret wrote: Did you try to add "-s lapw1" in your command line dedicated to the DFT+U calculation ? Remove any *dmat* files before. W. Le 26/06/2018 à 22:46, shaymlal dayananda a écrit : Dear William It is a error (messed up with another case.dayfile ) in the mail when I copy the day file to the mail, other than that , the problem and the errors are the same. Thank you Daya ___ Wien mailing list [ mailto:Wien@zeus.theochem.tuwien.ac.at | Wien@zeus.theochem.tuwien.ac.at ] [ http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien | http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien ] SEARCH the MAILING-LIST at: [ http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html | http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html ] -- William LAFARGUE-DIT-HAURET PhD Student -- Institut des Sciences Chimiques de Rennes UMR 6226 Equipe Chimie Théorique Inorganique Université de Rennes 1 - Campus de Beaulieu 35042 Rennes, France Building 10B, Desk 211 Phone: +33 (0)2.23.23.57.91 Email: [ mailto:william.lafargue-dit-hau...@univ-rennes1.fr | william.lafargue-dit-hau...@univ-rennes1.fr ] ___ Wien mailing list [ mailto:Wien@zeus.theochem.tuwien.ac.at | Wien@zeus.theochem.tuwien.ac.at ] [ http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien | http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien ] SEARCH the MAILING-LIST at: [ http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html | http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html ] ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Error in Hub-U calculation
Did you try to add "-s lapw1" in your command line dedicated to the DFT+U calculation ? Remove any *dmat* files before. W. Le 26/06/2018 à 22:46, shaymlal dayananda a écrit : Dear William It is a error (messed up with another case.dayfile ) in the mail when I copy the day file to the mail, other than that , the problem and the errors are the same. Thank you Daya ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html -- William LAFARGUE-DIT-HAURET PhD Student -- Institut des Sciences Chimiques de Rennes UMR 6226 Equipe Chimie Théorique Inorganique Université de Rennes 1 - Campus de Beaulieu 35042 Rennes, France Building 10B, Desk 211 Phone: +33 (0)2.23.23.57.91 Email: william.lafargue-dit-hau...@univ-rennes1.fr ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Error in Hub-U calculation
Dear William It is a error (messed up with another case.dayfile ) in the mail when I copy the day file to the mail, other than that , the problem and the errors are the same. Thank youDaya ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Error in Hub-U calculation
Hi, I noticed in your dayfile your filenames and directory name are inconsistent. -- case.dayfile is ; Calculating *U3O8-8.5-5000* in /scratch/WIEN2k17/*UB-8.5-5000* on gra144 with PID 28448 ... -- William Le 26/06/2018 à 21:29, shaymlal dayananda a écrit : Dear Prof. Blaha Thank you for your reply, I tried your instructions, but I am afraid that I am still getting the same error. This is what I did1. I copied case.energyup/dn as case.energyup_1/dn_12 copied case.dmatup_old and case.dmatup_old to case.dmatup and case.dmatup3. removed *.bro* ( But there were no files) And rerun the job script. But again I got the same error I got before. I noticed that now the energyup_1 file is empty again and case.enegydn_1 is missing. I repeated this in fresh directory, but nothing changes. My lapw with spin polarization runs without any problem. Only this hubard U inclusion gives troubles. Can this be an issue with my job script (attached in the previous reply) or installation? Thank you and I appreciate your valuable advice and support.I am re-sending this mail since I didn't get any reply yet. My apology about that. Daya ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html -- William LAFARGUE-DIT-HAURET PhD Student -- Institut des Sciences Chimiques de Rennes UMR 6226 Equipe Chimie Théorique Inorganique Université de Rennes 1 - Campus de Beaulieu 35042 Rennes, France Building 10B, Desk 211 Phone: +33 (0)2.23.23.57.91 Email: william.lafargue-dit-hau...@univ-rennes1.fr ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Error in hub-U calculation
Let me add something to that: If I remember correctly, if there is a problem with either the inorb or indm(c) files, I believe there is typically a forrtl error produced in the standard error/output. The forrtl error is usually helpful for narrow down what particular problem the input file has. That is if you compiled WIEN2k with -traceback. I believe one such error looks similar to [ https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg01885.html ]: forrtl: severe (24): end-of-file during read, unit 5, file /home/username/case.inorb Image PC Routine Line Source > orb 080BD261 Unknown Unknown Unknown ... Since it doesn't look like you attached the job script and provided the command you used for the job submission, there seems to be no way of knowing what your standard error/output files are called. As described in the slurm documentation [ https://slurm.schedmd.com/sbatch.html ], did you specify the standard error/output filename(s) using -e and -o flags, specify what they should be in your job script, or did you use the default (/By default both standard output and standard error are directed to a file of the name "slurm-%j.out", where the "%j" is replaced with the job allocation number. The file will be generated on the first node of the job allocation. Other than the batch script itself, Slurm does no movement of user files./)? On 6/19/2018 11:16 PM, Peter Blaha wrote: Probably you have wrong/inconsistent case.inorb and case.indm files (why do you have both indm and indmc ?? You should have just one file depending on inversion or no-inversion). Make sure they set the same density matrices. Eventually remove *dmat* and rerun. Am 20.06.2018 um 01:52 schrieb shaymlal dayananda: Dear developers and users We have recently got installed WIEN2k 17.1 and it uses slurm job submission manager. I am trying to do some test cases. Unfortunately our computer supporters could not make the job submission script with interface, so I do it manually. I tested the scf calculation with " runsp_lapw -NI -p -ec 0.0001 -cc 0.001 ". It completed without problem. However for adding hubbard-U , job crashes. I have created case.indm, case.indmc and case.inorb file and used the command "runsp_lapw -NI -p -dm -orb -ec 0.0001 -cc 0.001" in the job script. It failed with error "Error in Vorb" in uporb.error. I couldn't find any message in any relevant file. But it runs in single mode with "runsp_lapw -NI -p -dm -orb -ec 0.0001 -cc 0.001". Please help me to get corrected the parallel mode calculation. I appreciate any advice. I copied the dayfile and the uporb.def in my parallel mode calculation. case.dayfile is ; Calculating U3O8-8.5-5000 in /scratch/WIEN2k17/UB-8.5-5000 on gra144 with PID 28448 using WIEN2k_17.1 (Release 30/6/2017) in /home/software/Build_WIEN2k_17.1_M01/bin start (Tue Jun 19 19:42:02 EDT 2018) with lapw0 (140/99 to go) cycle 1 (Tue Jun 19 19:42:02 EDT 2018) (140/99 to go) > lapw0 -p (19:42:02) starting parallel lapw0 at Tue Jun 19 19:42:02 EDT 2018 .machine0 : processors running lapw0 in single mode 8.273u 0.309s 0:09.17 93.4% 0+0k 163816+12104io 654pf+0w > orb -up -p (19:42:12) 0.000u 0.037s 0:00.13 23.0% 0+0k 7248+32io 23pf+0w error: command /home/software/Build_WIEN2k_17.1_M01/bin/orb uporb.def failed > stop error UPORB.DEF: 5,'UB-8.5-5000.inorb', 'old', 'formatted',0 6,'UB-8.5-5000.outputorbup', 'unknown','formatted',0 9,'UB-8.5-5000.dmatdn', 'unknown','formatted',0 10,'UB-8.5-5000.dmatup', 'unknown','formatted',0 20,'UB-8.5-5000.struct', 'old', 'formatted',0 31,'UB-8.5-5000.br1orbup', 'unknown','unformatted',0 32,'UB-8.5-5000.br2orbup', 'unknown','unformatted',0 12,'UB-8.5-5000.vorbup', 'unknown','formatted',0 13,'UB-8.5-5000.vorbup_old', 'unknown','formatted',0 14,'UB-8.5-5000.energyup_1', 'unknown','formatted',0 18,'UB-8.5-5000.vspup', 'unknown','formatted',0 21,'UB-8.5-5000.scforbup', 'unknown','formatted',0 50,'UB-8.5-5000.eeceup', 'unknown','formatted',0 ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Error in hub-U calculation
Probably you have wrong/inconsistent case.inorb and case.indm files (why do you have both indm and indmc ?? You should have just one file depending on inversion or no-inversion). Make sure they set the same density matrices. Eventually remove *dmat* and rerun. Am 20.06.2018 um 01:52 schrieb shaymlal dayananda: Dear developers and users We have recently got installed WIEN2k 17.1 and it uses slurm job submission manager. I am trying to do some test cases. Unfortunately our computer supporters could not make the job submission script with interface, so I do it manually. I tested the scf calculation with " runsp_lapw -NI -p -ec 0.0001 -cc 0.001 ". It completed without problem. However for adding hubbard-U , job crashes. I have created case.indm, case.indmc and case.inorb file and used the command "runsp_lapw -NI -p -dm -orb -ec 0.0001 -cc 0.001" in the job script. It failed with error "Error in Vorb" in uporb.error. I couldn't find any message in any relevant file. But it runs in single mode with "runsp_lapw -NI -p -dm -orb -ec 0.0001 -cc 0.001". Please help me to get corrected the parallel mode calculation. I appreciate any advice. I copied the dayfile and the uporb.def in my parallel mode calculation. case.dayfile is ; Calculating U3O8-8.5-5000 in /scratch/WIEN2k17/UB-8.5-5000 on gra144 with PID 28448 using WIEN2k_17.1 (Release 30/6/2017) in /home/software/Build_WIEN2k_17.1_M01/bin start (Tue Jun 19 19:42:02 EDT 2018) with lapw0 (140/99 to go) cycle 1 (Tue Jun 19 19:42:02 EDT 2018) (140/99 to go) > lapw0 -p (19:42:02) starting parallel lapw0 at Tue Jun 19 19:42:02 EDT 2018 .machine0 : processors running lapw0 in single mode 8.273u 0.309s 0:09.17 93.4% 0+0k 163816+12104io 654pf+0w > orb -up -p (19:42:12) 0.000u 0.037s 0:00.13 23.0% 0+0k 7248+32io 23pf+0w error: command /home/software/Build_WIEN2k_17.1_M01/bin/orb uporb.def failed > stop error UPORB.DEF: 5,'UB-8.5-5000.inorb', 'old', 'formatted',0 6,'UB-8.5-5000.outputorbup', 'unknown','formatted',0 9,'UB-8.5-5000.dmatdn', 'unknown','formatted',0 10,'UB-8.5-5000.dmatup', 'unknown','formatted',0 20,'UB-8.5-5000.struct', 'old', 'formatted',0 31,'UB-8.5-5000.br1orbup', 'unknown','unformatted',0 32,'UB-8.5-5000.br2orbup', 'unknown','unformatted',0 12,'UB-8.5-5000.vorbup', 'unknown','formatted',0 13,'UB-8.5-5000.vorbup_old', 'unknown','formatted',0 14,'UB-8.5-5000.energyup_1', 'unknown','formatted',0 18,'UB-8.5-5000.vspup', 'unknown','formatted',0 21,'UB-8.5-5000.scforbup', 'unknown','formatted',0 50,'UB-8.5-5000.eeceup', 'unknown','formatted',0 ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html -- -- Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna Phone: +43-1-58801-165300 FAX: +43-1-58801-165982 Email: bl...@theochem.tuwien.ac.atWIEN2k: http://www.wien2k.at WWW: http://www.imc.tuwien.ac.at/tc_blaha- ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
