- mp_world and world_comm should be used only in MPI initialization,
not directly in codes - replaced by mp_images and intra_image_comm
(except in a few cases whose usage is unclear to me)
- mp_global should be used only to provide initialization routines
mp_startup, mp_global_end
- specific mp_* modules should be used for parallelization level on *
(* = images, pools, bands, orhtoh, ...)
Nothing actually changes (I hope) but things will look cleaner if the module
for the required parallelization level is USEd instead of a generic one that
does too meny things at the same time. The final goal of this operation is to
better disentangle the various parallelization levels and to have a better
organized, more granular MPI initialization that does not introduces artificial
and meaningless dependencies. Not done for: PHonon, CPV, TDDFPT, GWW, EPW
uneeded relinking; the mysterious "tldeps" machinery modified accordingly;
added explicit dependencies to GWW targets to prevent (once again) problems
with make -j (seems to work now). make.depend updated (once again).
impossible clashes with LIBS as defined by the OS or as used by the configure's
of external libraries (e.g. FoX, for which one such problem has been reported)
Since github and gitlab just distribute a snapshot of the git repository,
these files must be in the repository as well. NOTA BENE:
All changes MUST GO TO *.tex or *.def FILES, NOT to these files!
Also note that the "make doc" machinery produces also *.xml files and
latex2html converted html versions of latex files. I don't think they belong
to the repository (latex files should maybe be replaced by md files)
Makefiles modified accordingly.
I am not at all happy about this "solution": unneeded dependencies should be
avoided, but this would need to change the way things are deallocated at the
end of a run (clean_pw does too many things at the same time)
after the proliferarion of modules and libraries, I have collected into a
single variable BASEMOD_FLAGS, defined in make.inc, the list of directories
for modules, to be included by all codes. IMPORTANT NOTICE: re-run "configure",
the new Makefiles work in conjunction with the new make.inc
In real space processors are organized in a 2D pattern.
Each processor owns data from a sub-set of Z-planes and a sub-set of Y-planes.
In reciprocal space each processor owns Z-columns that belong to a sub set of
X-values. This allows to split the processors in two sets for communication
in the YZ and XY planes.
In alternative, if the situation allows for it, a task group paralelization is used
(with ntg=nyfft) where complete XY planes of ntg wavefunctions are collected and Fourier
trasnformed in G space by different task-groups. This is preferable to the Z-proc + Y-proc
paralleization if task group can be used because a smaller number of larger ammounts of
data are transferred. Hence three types of fft are implemented:
!
!! ... isgn = +-1 : parallel 3d fft for rho and for the potential
!
!! ... isgn = +-2 : parallel 3d fft for wavefunctions
!
!! ... isgn = +-3 : parallel 3d fft for wavefunctions with task group
!
!! ... isgn = + : G-space to R-space, output = \sum_G f(G)exp(+iG*R)
!! ... fft along z using pencils (cft_1z)
!! ... transpose across nodes (fft_scatter_yz)
!! ... fft along y using pencils (cft_1y)
!! ... transpose across nodes (fft_scatter_xy)
!! ... fft along x using pencils (cft_1x)
!
!! ... isgn = - : R-space to G-space, output = \int_R f(R)exp(-iG*R)/Omega
!! ... fft along x using pencils (cft_1x)
!! ... transpose across nodes (fft_scatter_xy)
!! ... fft along y using pencils (cft_1y)
!! ... transpose across nodes (fft_scatter_yz)
!! ... fft along z using pencils (cft_1z)
!
! If task_group_fft_is_active the FFT acts on a number of wfcs equal to
! dfft%nproc2, the number of Y-sections in which a plane is divided.
! Data are reshuffled by the fft_scatter_tg routine so that each of the
! dfft%nproc2 subgroups (made by dfft%nproc3 procs) deals with whole planes
! of a single wavefunciton.
!
fft_type module heavily modified, a number of variables renamed with more intuitive names
(at least to me), a number of more variables introduced for the Y-proc parallelization.
Task_group module made void. task_group management is now reduced to the logical component
fft_desc%have_task_groups of fft_type_descriptor type variable fft_desc.
In term of interfaces, the 'easy' calling sequences are
SUBROUTINE invfft/fwfft( grid_type, f, dfft, howmany )
!! where:
!!
!! **grid_type = 'Dense'** :
!! inverse/direct fourier transform of potentials and charge density f
!! on the dense grid (dfftp). On output, f is overwritten
!!
!! **grid_type = 'Smooth'** :
!! inverse/direct fourier transform of potentials and charge density f
!! on the smooth grid (dffts). On output, f is overwritten
!!
!! **grid_type = 'Wave'** :
!! inverse/direct fourier transform of wave functions f
!! on the smooth grid (dffts). On output, f is overwritten
!!
!! **grid_type = 'tgWave'** :
!! inverse/direct fourier transform of wave functions f with task group
!! on the smooth grid (dffts). On output, f is overwritten
!!
!! **grid_type = 'Custom'** :
!! inverse/direct fourier transform of potentials and charge density f
!! on a custom grid (dfft_exx). On output, f is overwritten
!!
!! **grid_type = 'CustomWave'** :
!! inverse/direct fourier transform of wave functions f
!! on a custom grid (dfft_exx). On output, f is overwritten
!!
!! **dfft = FFT descriptor**, IMPORTANT NOTICE: grid is specified only by dfft.
!! No check is performed on the correspondence between dfft and grid_type.
!! grid_type is now used only to distinguish cases 'Wave' / 'CustomWave'
!! from all other cases
Many more files modified.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@13676 c92efa57-630b-4861-b058-cf58834340f0
KS_Solvers/CG, KS_Solvers/Davidson, KS_Solvers/Davidson_RCI.
Two are currently used by QE, the third one implements the Davidson
diagonalization within the Reverse Communication Interface paradigm,
courtesy of Micael Oliveira.
KS_Solvers routines depend only on lower level libraries, notably UtilXlib,
LAXlib, (SCA)LAPACK, and BLAS.
reorganization can be improved. For instance some duplicated routines like
cdiaghg and rdiaghg could/should be moved in LAXlib. This could reduce the need
to include KS_Solvers directories in the link step of many codes.
Minimal changes to calling sequence have been made, essentially just adding
h_psi,s_psi,g_psi and h_1psi,s_1psi routines names as arguments (with a
specific calling sequence ihardcode inside the routines that agree with PWSCF one).
This could be avoided adopting the RCI paradigm.
Compiled in serial and parallel, 177/182 pw tests passed (3 that were failing
even before on my laptop pw-berry, pw-langevin, pw-pawatom + 2 unknown==not tested),
12 /17 cp tests passed (some o2-us-para-pbe-X fail but the same was for the
original version)
I assume the modified calling procedure is working and the problem lies somewhere else.
Randomly tested some examples in pw, ph, pwcond and it seams to work.
Please report any problem.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@13644 c92efa57-630b-4861-b058-cf58834340f0
basic operations: error handling, timing clocks, interfaces to basic mpi
calls, find free units...
These routines are moved from Modules and dependencies to other modules
are removed.
MANY files are updated to comply with the move.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@13629 c92efa57-630b-4861-b058-cf58834340f0
US variable qq renamed qq_nt and a new variable qq_na added
because in real space the integral may depend (slightly) on
the atomic position and an atomic value is needed to compute
exactly normalizable wfc.
Whenever realspace tricks are not used qq_nt is used.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@13604 c92efa57-630b-4861-b058-cf58834340f0
to compute many FFTs at the same time, particularly usefull for EXX
but could be usefule for many linear response code as well
(for the time being implemented only for DFTI and internal FFTW,
should be trivial to extend other drivers)
- more clean-ups
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@12815 c92efa57-630b-4861-b058-cf58834340f0
prevent trouble with OS-X. May or may not work (it won't unless configure
is updated: please somebody with v.2.63 of autoconf do it), may turn out to
be obsolete anyway.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@12364 c92efa57-630b-4861-b058-cf58834340f0
moment to get rid of flib/, whose usefulness is far from obvious. The content
of flib/ is now in Modules/. Many makefiles updated and little more.
Packages using QE routines should just remove links to flib/flib.a.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@12072 c92efa57-630b-4861-b058-cf58834340f0
Paolo Giannozzi