[Wien] building Wien2k against hpmpi
Hi, Is there anybody running Wien2k (09_01) with hpmpi (2.3.1) on Linux (SLES10sp2) and Intel x86_64 processors ? When using this configuration, I got a crash in lapw2para (lap2para -up uplapw2.def). Lapw0 & lapw1 completed successfully, then lapw2 is saying "LAPW2 - FERMI; weights written" then lapw2 is crashing : LAPW1 END LAPW1 END LAPW1 END LAPW1 END LAPW2 - FERMI; weighs written mpirun: Info: plugin plugin002_007_001.so being skipped due to setting of mpirun: HPMPI_IGNORE_002_007_001 in environment mpirun: Info: plugin plugin002_007_001.so being skipped due to setting of mpirun: HPMPI_IGNORE_002_007_001 in environment mpirun: Info: plugin plugin002_007_001.so being skipped due to setting of mpirun: HPMPI_IGNORE_002_007_001 in environment mpirun: Info: plugin plugin002_007_001.so being skipped due to setting of mpirun: HPMPI_IGNORE_002_007_001 in environment mpirun: Info: plugin plugin002_007_001.so being skipped due to setting of mpirun: HPMPI_IGNORE_002_007_001 in environment MPI Application rank 4 killed before MPI_Finalize() with signal 11 forrtl: error (78): process killed (SIGTERM) Image PCRoutineLine Source libpthread.so.02B7FAB160C00 Unknown Unknown Unknown libpthread.so.02B7FAB15F880 Unknown Unknown Unknown lapw2_mpi 004A05EF Unknown Unknown Unknown lapw2_mpi 004A008B Unknown Unknown Unknown lapw2_mpi 004C9127 Unknown Unknown Unknown lapw2_mpi 00481BF1 recpr_219 recpr.f lapw2_mpi 0047EC4C recfil_ 106 recfil.f lapw2_mpi 0046C171 MAIN__249 lapw2_tmp_.F lapw2_mpi 00403EBC Unknown Unknown Unknown libc.so.6 2B7FAB28A184 Unknown Unknown Unknown lapw2_mpi 00403DE9 Unknown Unknown Unknown MPI Application rank 42 killed before MPI_Finalize() with signal 11 MPI Application rank 56 killed before MPI_Finalize() with signal 11 MPI Application rank 23 killed before MPI_Finalize() with signal 11 Thanks, Bruno PS : I'm rebuilding Blacs & Scalapack against hpmpi (2.3.1) with Intel Fortran 11.0.074 & mkl 10.1.1.1.019, Wien2k is built in dynamic mode . -- Bruno Monnet HP Invent SCI EMEA Competency Center CMU : www.hp.com/go/cmu 5 av R. Chanas 38053 Grenoble CEDEX 9 France mailto: bruno.monnet at hp.com <mailto:bruno.monnet at hp.com> Phone: Mobile: Fax: +33 (0)4.76.14.17.84 +33 (0)6.72.99.27.32 +33 (0)1.76.72.87.02 -- next part -- An HTML attachment was scrubbed... URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20090703/f73651bf/attachment.htm> -- next part -- A non-text attachment was scrubbed... Name: image001.jpg Type: image/jpeg Size: 1960 bytes Desc: not available URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20090703/f73651bf/attachment.jpg>
[Wien] A question about unit of electric field in in0
Dear Prof. Blaha, According to WIEN2k v09.1 and v08.3, if the EFIELD variable in case.in0 is set to 1.0, then difference of total potential (I generated it using lapw5 ATU VAL options) with EFIELD=1.0 and without EFIELD line, shows up a zig-zag variation in a unit cell from 0.5 to -0.5 (for analytic triangular ramp). My case.in0 is asking 40 Fourier coeficients for analytical triangular ramp: --- TOT 13(5...CA-LDA, 13...PBE-GGA, 11...WC-GGA) R2V IFFT (R2V) 30 30 2431.00min IFFT-parameters, enhancement factor 40 1.0 --- Am I right that the electric field, let say at z=0.25c, is then equal to 2 EFIELD/c (where c is the lattice constant along z-direction) ? Does it not mean that in your previous discussion from May 23, 2008, please see below, the value of the electric field is two times the 173 kV/nm and not four times? What have you meant by the external potential of 100mRy in your PRB 63, 165205 (2001)? With the best regards, Martin Gmitra *Fri May 23 08:29:30 CEST 2008* You must multiply by 4 ! (because field varies from +100 to -100 and since it is a zig-zag, it varies in half the period) zhao schrieb: >* Dear Prof. P. Blaha, *>* *>* In your paper, PRB-63-165205(2001), you say that external potential *>* 100mRy corresponds to 700 kV/mm. When I tried to get this unit *>* transformation, I was confused by something. The following is my *>* problem. *>* -- *>* c=8*sqrt(3)*a0 *>* a0=5.65 Angs *>* Efield= 0.1Ry/c=0.1*13.6V/(8*sqrt(3)*a0)= 173 kV/mm *>* -- *>* *>* I cannot get 700 kV/mm. Can you tell me what's wrong with myself? *>* Thank you very much. *>* *>* *>* *>* Best wishes, *>* *>* Sincerely, *>* Yong-Hong Zhao* -- next part -- An HTML attachment was scrubbed... URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20090703/49d3fbcb/attachment.htm>
