Commit Graph

5320 Commits

Author SHA1 Message Date
giannozz beb533efe3 Fixed recently introduced bug in force constant calculation
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5252 c92efa57-630b-4861-b058-cf58834340f0
2008-10-30 08:03:05 +00:00
ceresoli 1be9791025 I changed this because I've got an 'array out of bounds' in some calculations.
Can someone please check if my fix is ok? (D.C.)


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5251 c92efa57-630b-4861-b058-cf58834340f0
2008-10-28 18:54:51 +00:00
cavazzon 454822a037 - New verbosity rules:
iprint defines how often quantities are written to trajectory files,
  whereas the same quantities are printed on standard output
  depending on the value of the verbosity keyword.
  With verbosity = 'default' or verbosity = 'low'
  quantities are replicated on stdout every 10*iprint steps,
  with verbosity = 'minimal' every 100*iprint steps and
  with verbosity = 'normal' every iprint steps

WARNING: THIS WILL CHANGE THE OUTPUT OF THE EXAMPLES


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5250 c92efa57-630b-4861-b058-cf58834340f0
2008-10-28 17:38:01 +00:00
giannozz 0f392bc7de Bug for spin-polarized case (noticed by Ding Yi)
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5249 c92efa57-630b-4861-b058-cf58834340f0
2008-10-28 07:46:11 +00:00
giannozz ad8d8a05ac Yet another compilation problem in serial case
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5248 c92efa57-630b-4861-b058-cf58834340f0
2008-10-28 07:44:52 +00:00
giannozz 0526f6ac21 More configure tweakings, preventing env. variable F77 to make trouble
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5247 c92efa57-630b-4861-b058-cf58834340f0
2008-10-27 11:04:24 +00:00
giannozz c1d4096628 Detection of ffts in acml now really seems to work
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5246 c92efa57-630b-4861-b058-cf58834340f0
2008-10-27 10:12:08 +00:00
giannozz 77f1aa4b1f More configure tweakings, removal of obsolete stuff. Blas, Lapack, fft from
ACML should be recognized when using PGI and pathscale, intel not yet.


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5245 c92efa57-630b-4861-b058-cf58834340f0
2008-10-27 09:41:01 +00:00
giannozz 3446c2f5e3 #ifdef __SCALAPACK moved around so that serial compilation works.
make.depend's updated


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5244 c92efa57-630b-4861-b058-cf58834340f0
2008-10-27 08:33:19 +00:00
cavazzon d01ce581f8 - Scalapack interface
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5243 c92efa57-630b-4861-b058-cf58834340f0
2008-10-26 11:08:28 +00:00
cavazzon 488d34b986 - SCALAPACK interface for real matrix too
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5242 c92efa57-630b-4861-b058-cf58834340f0
2008-10-26 09:28:59 +00:00
cavazzon c6324310e2 completed the interface with SCALAPACK, now is working but is not fully tested.
Scalapack is used to diagonalize block distributed matrixes (cdiaghg),
great speed-up are possible for large matrixes ( > 1000 ) when
using a large number of processors ( > 512 ) .
If you want to test scalapack you have to compile adding -D__SCALAPACK
to DFLAGS in make.sys and you have to modify the LAPACK_LIBS varaible
like in:
SCALAPACK_LIBS = /cineca/prod/libraries/SCALAPACK/1.8.0/openmpi--1.2.5--intel--10.1/libscalapack.a
BLACS_LIBS     = /cineca/prod/libraries/BLACS/1.1/openmpi--1.2.5--intel--10.1/libblacs.a
BLACS_INI      = /cineca/prod/libraries/BLACS/1.1/openmpi--1.2.5--intel--10.1/libblacsF77init.a
LAPACK_LIBS    = $(SCALAPACK_LIBS) $(BLACS_LIBS) $(BLACS_INI) $(BLACS_LIBS) /cineca/prod/acml/4.1.0/ifort64/lib/libacml.a


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5241 c92efa57-630b-4861-b058-cf58834340f0
2008-10-26 07:39:53 +00:00
giannozz 4e8eb847ac Ooops..
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5240 c92efa57-630b-4861-b058-cf58834340f0
2008-10-24 14:24:22 +00:00
giannozz 8d339650a3 Minor changes to installation and its documentation
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5239 c92efa57-630b-4861-b058-cf58834340f0
2008-10-24 12:41:05 +00:00
paulatto b622dcc6c0 Some of the files contained dos-style line terminations that can confuse
qe compiled with xlf.

LP


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5238 c92efa57-630b-4861-b058-cf58834340f0
2008-10-24 09:46:28 +00:00
dalcorso 2060d5cdc0 Added the possibility to read as input k-points only the first and
last point of a line. The weight of the first k-point
is the number of points along the line. (use the keyword tpiba_b or
crystal_b to use this feature).


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5237 c92efa57-630b-4861-b058-cf58834340f0
2008-10-23 15:07:39 +00:00
dalcorso 39a065f51d Several improvements, problems/bug fixes needed for the grid.
(Thanks to R. Di Meo for pointing out some of the problems).


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5236 c92efa57-630b-4861-b058-cf58834340f0
2008-10-23 14:47:04 +00:00
smogunov 238c8882c4 Added the possibility to calculate and plot the scattering states (A. Smogunov).
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5235 c92efa57-630b-4861-b058-cf58834340f0
2008-10-23 13:47:43 +00:00
dalcorso 3fc301b28e Bug fix: the ph.x code was not working when wf_collect=.true. and lnscf=.true.
or ldisp=.true..


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5234 c92efa57-630b-4861-b058-cf58834340f0
2008-10-23 07:04:43 +00:00
ceresoli 58e8d82d3f `#endif' instead of `#end if' (D.C.)
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5233 c92efa57-630b-4861-b058-cf58834340f0
2008-10-22 20:13:29 +00:00
cavazzon b1b1e64797 - adding dimension and mapping functions required to interface QE with SCALAPACK
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5232 c92efa57-630b-4861-b058-cf58834340f0
2008-10-22 12:16:27 +00:00
cavazzon 6d0aae0f27 - adding variables and setup required to interface QE with SCALAPACK
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5231 c92efa57-630b-4861-b058-cf58834340f0
2008-10-22 12:14:50 +00:00
fratesi 2dbc1bc481 Reference for NIST results corrected (PRB->PRA)
Guido F


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5230 c92efa57-630b-4861-b058-cf58834340f0
2008-10-20 09:25:52 +00:00
giannozz aa2daf4052 Incorrect commit reverted
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5229 c92efa57-630b-4861-b058-cf58834340f0
2008-10-18 14:43:02 +00:00
giannozz 0b8cbf3b75 Configure modified so that it recognizes FFTW3 and FFT libraries from ACML.
Mostly untested. Support for external fftw v.2 removed.


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5228 c92efa57-630b-4861-b058-cf58834340f0
2008-10-18 14:40:51 +00:00
dalcorso ba4cbb2a3d Corrected another problem introduced with recent commit.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5227 c92efa57-630b-4861-b058-cf58834340f0
2008-10-18 14:17:28 +00:00
dalcorso 78945be233 Corrected some problems when start_irr=0 and last_irr=0.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5226 c92efa57-630b-4861-b058-cf58834340f0
2008-10-18 14:11:44 +00:00
giannozz 53928387b0 Documentation updated
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5224 c92efa57-630b-4861-b058-cf58834340f0
2008-10-17 16:04:01 +00:00
giannozz 652a7553d9 Minor changes to initial printout
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5223 c92efa57-630b-4861-b058-cf58834340f0
2008-10-17 16:02:57 +00:00
dalcorso 8efedc981f Some speed-up in the final diagonalization when different representations
are calculated in different machines.


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5222 c92efa57-630b-4861-b058-cf58834340f0
2008-10-10 14:50:07 +00:00
giannozz a8c7b33e60 Units of q vectors explicitly specified
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5221 c92efa57-630b-4861-b058-cf58834340f0
2008-10-10 09:52:00 +00:00
paulatto 4aaca9ff22 Printout format fixed and clarified
LP


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5220 c92efa57-630b-4861-b058-cf58834340f0
2008-10-10 08:35:24 +00:00
giannozz 7bcafa1dba array qv allocated in newd wasn't deallocated in all cases,
leading to either a crash or a memory leak (Ralph)


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5219 c92efa57-630b-4861-b058-cf58834340f0
2008-10-07 16:00:21 +00:00
paulatto f50125be05 removed unnecessary used module
LP


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5218 c92efa57-630b-4861-b058-cf58834340f0
2008-10-06 11:49:58 +00:00
cavazzon b5ece34b51 - Task groups bug fix, array tg_rho was not cleared at every k point cycle.
This was causing problems with some combinations of "-npool" and "-ntg".


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5217 c92efa57-630b-4861-b058-cf58834340f0
2008-10-05 07:48:08 +00:00
ceresoli 6928468867 Added missing call to setup(). (D.C.)
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5216 c92efa57-630b-4861-b058-cf58834340f0
2008-10-02 19:50:03 +00:00
dalcorso 1f86eab25b Small bug fix. After the recent changes to the symmetry routines the phonon code
in some cases wrote strange output such as: 48 Sym.Ops. (no inversion).


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5215 c92efa57-630b-4861-b058-cf58834340f0
2008-09-30 09:16:44 +00:00
giannozz 2d0ec808c3 Input variables were used in a declaration before they were defined - some
compilers complain, some don't


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5214 c92efa57-630b-4861-b058-cf58834340f0
2008-09-25 08:32:32 +00:00
dalcorso 693a3aac1b Small change.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5213 c92efa57-630b-4861-b058-cf58834340f0
2008-09-24 13:14:14 +00:00
dalcorso 869b76fdab Small output change in the case iverbosity=1.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5212 c92efa57-630b-4861-b058-cf58834340f0
2008-09-24 13:08:23 +00:00
giannozz 1404b5f8cd More pwcom splitting into modules
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5211 c92efa57-630b-4861-b058-cf58834340f0
2008-09-24 11:50:31 +00:00
smogunov 6defd54dcf A bug with some array bounds fixed (A. Smogunov)
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5210 c92efa57-630b-4861-b058-cf58834340f0
2008-09-24 11:30:13 +00:00
giannozz 37d2d944ff Forgot to modify f77 to f90 format as weel
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5209 c92efa57-630b-4861-b058-cf58834340f0
2008-09-23 09:19:31 +00:00
giannozz 1fc3e74d7f Multigrid *.f files should be f77 files but actually they contain some
elements of f90 syntax, that makes them not compilable by f77. I renamed
a single file containing allocatable arrays to *.f90, removed array syntax
in a few places, so that now the naming convention is consistent with the
rest of q-e. There are still a few problems left: routines cgmg and fbound00
are called with a variable number of arguments (this is incorrect by the way),
routine random_number shouldn't be used, headers should be updated to reflect
the cahnges to the orginal code


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5208 c92efa57-630b-4861-b058-cf58834340f0
2008-09-23 09:10:09 +00:00
giannozz 66b732747a Minor cleanup
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5207 c92efa57-630b-4861-b058-cf58834340f0
2008-09-22 17:30:16 +00:00
dalcorso e23e3527cc Bug fix: problem with the generation of the atomic wavefunctions in the
projwfc code when a scalar relativistic PP is used with lspinorb=.true..


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5206 c92efa57-630b-4861-b058-cf58834340f0
2008-09-22 09:28:44 +00:00
giannozz 34b5009948 Better workaround for clean_pw
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5205 c92efa57-630b-4861-b058-cf58834340f0
2008-09-19 09:53:11 +00:00
giannozz f946339d97 Some undesired changes slipped into yesterday's commit.
Workaround for crash in pwcond after recenbt changes to phonon symmetry:
a more definitive solution is needed for clean_pw, though


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5204 c92efa57-630b-4861-b058-cf58834340f0
2008-09-19 08:46:31 +00:00
giannozz b14ae22a3e Removed q-point from PWscf (variable xqq), removed phonon-related variables
from xml file (beware unexpected side effects in phonon restart).


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5203 c92efa57-630b-4861-b058-cf58834340f0
2008-09-18 20:27:46 +00:00
dalcorso 33c05541cd Problem fix: zue + gamma_gamma tricks was not implemented. Example09
updated to calculate also the response to the electric field.


git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@5202 c92efa57-630b-4861-b058-cf58834340f0
2008-09-18 13:21:43 +00:00