lcpq / quantum_package Goto Github PK

View Code? Open in Web Editor NEW

44.0 19.0 38.0 39.62 MB

Set of quantum chemistry programs and libraries

Home Page: https://quantumpackage.github.io/qp2/

License: GNU General Public License v2.0

Makefile 0.09% OCaml 9.38% Shell 0.85% Python 5.47% Fortran 84.16% C 0.06%

fortran ocaml zmq amazing computational-chemistry quantum quantum-chemistry-programs chemistry wavefunction dft

quantum_package's People

Contributors

Stargazers

Watchers

quantum_package's Issues

Update readme

For now update_readme.py can only work if the module have a IRP_MAN and a tag.
But we never use irp on module who have not binarys.

The main idea is to alow update_reade to udapte all the descendant module if needed.

MD5 key for MOs

Reading MO integrals should check the MD5 key of the MOs to be sure that they correspond.

All the python scripts need to use `qp_path`

All the global variable relatif to path are now stored in qp_path.
We need to make the changen in the script.

EZFIO should compiled with --openmp

STOP COMMIT IN THIS REPO

Plz, stop commit in this repo. 😠 👿 (if anyone ever do this)
The new idea:

Make a Pull Request
Wait for the Travis Test to pass (https://travis-ci.org/LCPQ/quantum_package/)
When you have a green check, accept the pull request 👍 😄

dot makes an error if graphviz is not installed

[9/30] Generate Png representtion of the Tree Dependencies of Electrons
FAILED: cd /mnt/fhgfs/scemama/quantum_package/src/Electrons ; module_handler.py create_png
Couldn't import dot_parser, loading of dot files will not be possible.
Traceback (most recent call last):
  File "/mnt/fhgfs/scemama/quantum_package/scripts/module/module_handler.py", line 238, in <module>
    m.create_png(l_module)
  File "/mnt/fhgfs/scemama/quantum_package/scripts/module/module_handler.py", line 211, in create_png
    graph.write_png(path)
  File "/mnt/fhgfs/scemama/quantum_package/scripts/utility/pydot.py", line 1903, in <lambda>
    prog=prog))
  File "/mnt/fhgfs/scemama/quantum_package/scripts/utility/pydot.py", line 1999, in write
    dot_fd.write(self.create(prog, format))
  File "/mnt/fhgfs/scemama/quantum_package/scripts/utility/pydot.py", line 2039, in create
    'GraphViz\'s executables not found')
pydot.InvocationException: GraphViz's executables not found

`qp_create_ezfio_from_xyz` ask for to much atoms

qp_create_ezfio_from_xyz create a tmpfile who contains ALL the element in the basis-set.

And if the Basis only have He, it will bug and say

Uncaught exception:

  (Failure "Element C not found")

Maybe only get the element needed. (Aka call EMSL with --atom arg)

Checking irpf90...
Traceback (most recent call last):
File "./configure", line 501, in
l_installed, l_install_descendant = checking(d_dependency)
File "./configure", line 260, in checking
r = check_availability(i)
File "./configure", line 211, in check_availability
version = check_output("irpf90 -v".split()).strip()
File "./configure", line 168, in check_output
stderr=subprocess.PIPE, _popenargs, *_kwargs)
File "/usr/lib64/python2.6/subprocess.py", line 642, in init
errread, errwrite)
File "/usr/lib64/python2.6/subprocess.py", line 1234, in _execute_child
raise child_exception
OSError: [Errno 8] Exec format error

Exit Code in ezfio

We need to have a exit code on the ezfio for checking if the calcul fail or not.

Something like a file status with "Running" or "Finish" inside. Or a boolean.

.gitignore seems broken

make veryclean ; make # Broken with EZFIO.cfg

cd src
make veryclean
make

...

IRPF90_temp/MonoInts/pot_ao_ints.irp.F90(102): error #6404: This name does not have a type, and must have an explicit type.   [DO_PSEUDO]
  if (do_pseudo.eqv..TRUE.) then                       ! MonoInts/pot_ao_ints.irp.f:  13
------^
compilation aborted for IRPF90_temp/MonoInts/pot_ao_ints.irp.F90 (code 1)
make[1]: *** [IRPF90_temp/MonoInts/pot_ao_ints.irp.o] Error 1
make[1]: *** Waiting for unfinished jobs....
make[1]: Leaving directory `/home/scemama/quantum_package/src/Bielec_integrals'
-------------------- Error --------------------

Build failed for module Bielec_integrals

-----------------------------------------------
make: *** [default] Error 255

Because MonoInts/EZFIO.cfg is does not rebuild do_pseudo in Bielec_integrals.
All EZFIO.cfg files should be transformed to *.irp.f files before.

unknown config file in configure

configure should give a clean error message with wrong ./config.* file.

$ ./configure ./config/ifort --production
Traceback (most recent call last):
  File "/ccc/work/cont003/gen1738/scemama/quantum_package/scripts/compilation/qp_create_ninja.py", line 811, in <module>
    l_string = ninja_create_env_variable(pwd_config_file)
  File "/ccc/work/cont003/gen1738/scemama/quantum_package/scripts/compilation/qp_create_ninja.py", line 88, in ninja_create_env_variable
    flag))
  File "/ccc/work/cont003/gen1738/scemama/quantum_package/scripts/compilation/read_compilation_cfg.py", line 22, in get_compilation_option
    if flag_name == "FC" and config.getboolean("OPTION","CACHE"):
  File "/usr/lib64/python2.6/ConfigParser.py", line 349, in getboolean
    v = self.get(section, option)
  File "/usr/lib64/python2.6/ConfigParser.py", line 532, in get
    raise NoSectionError(section)
ConfigParser.NoSectionError: No section: 'OPTION'
Creating build.ninja...       
Traceback (most recent call last):
  File "./configure", line 499, in <module>
    create_ninja_and_rc(l_installed)
  File "./configure", line 469, in create_ninja_and_rc
    subprocess.check_call(" ".join(l), shell=True)
  File "/usr/lib64/python2.6/subprocess.py", line 505, in check_call
    raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '/ccc/work/cont003/gen1738/scemama/quantum_package/scripts/compilation/qp_create_ninja.py create config/ifort --production' returned non-zero exit status 1

Documentation of EZFIO in Fortran

A python script ezfio_interface.py will read a DEPENDENCY config file with ConfigParser of the form (for example)

[n_det]
doc:        Number of states to consider
ezfio_name: det_num
ezfio_type: integer
ezfio_dir:  determinants

and will create the *.ezfio_config file, and the ezfio_interface.irp.f file.

from ezfio_with_default import EZFIO_Provider can be used to generate ezfio_interface.irp.f.

`/irp_checkpoint.irp.o` Missing

Witch the new IRPF90 version (1.6.1), the EZFIO folder in QP don't compile. Throw this error :

`bash -c "[[ -e ../libezfio.a ]] && rm ../lib/libezfio.a"
make[1]: [../lib/libezfio.a] Error 1 (ignored)
ar cru ../lib/libezfio.a IRPF90_temp/irp_stack.irp.o  IRPF90_temp/libezfio_error.irp.o IRPF90_temp/libezfio_error.irp.module.o IRPF90_temp/libezfio_file.irp.o IRPF90_temp/libezfio_file.irp.module.o IRPF90_temp/libezfio_util.irp.o IRPF90_temp/libezfio_util.irp.module.o IRPF90_temp/libezfio_groups.irp.o IRPF90_temp/libezfio_groups.irp.module.o  IRPF90_temp/irp_touches.irp.o IRPF90_temp/irp_checkpoint.irp.o 
ar: IRPF90_temp/irp_checkpoint.irp.o: No such file or directory
make[1]: *** [../lib/libezfio.a] Error 1
make[1]: Leaving directory `/tmp/quantum_package/EZFIO/src'
make: *** [default] Error 2

HOT FIX

In ${QPACKAGE_ROOT}/scripts/install_irpf90.sh change

URL="https://github.com/LCPQ/${BASE}/archive/master.tar.gz"

into

URL="https://github.com/TApplencourt/${BASE}/archive/master.tar.gz"

Delete ${QPACKAGE_ROOT}/irp90, and run setup_environement.sh

Install patch

Bug OCAML in input_ao_basis.ml

ocamlbuild -j 0 -syntax camlp4o -cflags "-g -warn-error A" -lflags "-g -warn-error A" -ocamlopt ocamlc.opt test_input.byte  -use-ocamlfind  
+ ocamlfind ocamlc -c -g -warn-error A -thread -package core,sexplib.syntax,cryptokit -syntax camlp4o -o Input_ao_basis.cmo Input_ao_basis.ml
File "Input_ao_basis.ml", line 23, characters 24-39:
Error: Unbound module AO_basis_name
Command exited with code 2.

Bitmasks for input

Create bitmasks from input file for orbital classes.

test.irp.f

When a test.irp.f file is present the directory doesn't compile.

Buildbot

Create a buildbot system to run tests

CISD_SC2 Selection

qp_run cid_sc2_selection

CISD SC2
========
 provide_ao_basis_ao_num: Error: Provider in an openMP section

Readme.rst conflicts

Automatic build of Readme.rst should be done only for the LCPQ branch. Forks should not automatically update Readme.rst because there are lots of conflicts durin merges.

REPO=$( git config --get remote.origin.url | cut -d ':' -f 2 )

should be LCPQ/quantum_package.git . If not, update_readme.py should do nothing.

L symmetry in n-31g basis sets

5
methane molecule 
C        0.000000        0.000000        0.000000
H        0.000000        0.000000        1.089000
H        1.026719        0.000000       -0.363000
H       -0.513360       -0.889165       -0.363000
H       -0.513360        0.889165       -0.363000

$ ./qp_create_ezfio_from_xyz roger.xyz -b 3-21++g -m 3

Uncaught exception:

  Gto.GTO_Read_Failure("  1      3.6649800             -0.3958970              0.2364600        ")

Raised at file "Gto.ml", line 62, characters 19-49
Called from file "Gto.ml", line 64, characters 5-20
Called from file "Basis.ml", line 10, characters 16-39
Called from file "Basis.ml", line 11, characters 6-37
Called from file "lib/core_list.ml", line 404, characters 13-17
Called from file "lib/command.ml", line 1678, characters 6-68
Called from file "lib/exn.ml", line 80, characters 6-10

update_README.py is broken

Current output file

Create a global variable for the current output file

Ezfio -> API -> Ocaml -> Fortran

List of data needed by Ocaml :

(Name of the provider,Function to call)
All the provider who is a Option
All the integer default value ?

👍

Check scripts/ezfio_interface/convert_ezfio.sh

Integral AO

It seem to have a bug.
The calculus of integrals for the Ti atom in cc-pvdz with BFD pseudo take a long long time

Fortran should let ocaml check inputs

Remove all EZFIO input checking in Fortran

Add doc to qp_convert.py and rename it

Hi @eginer @scemama,

I don't know what qp_convert.py doing. I guess is convert somme file who can be recognize by result_file utility to EZFIO, but i not sure.

Can we set a explicit name to this script and add a few documentation? Maybe a qp_convert_output_to_ezfio.py. (Cause we a have a qp_convert_ezfio_v1_to_v2.sh so it is a little tedious to have a qp_convert.py)

Long&explicit name for conversion script are relevant i think.

Move it to the the /script/ezfio_interface folder seem logical too. (All the stuff related to the ezfio need to go inside).

Have a nice day,

PS : I can do the change if you want. I was just asking, because is not my script : )

Thomas

Hartree-Fock

Uninstall Module

We need to remove the ocaml/Input_$MODULE.ml file is needed.

The ezfio seem to be not rebuild as well.

Initial Guess

HCore
Huckel

Default not handle by qp_edit

Not the case for:

Mo_tot_num
Mo_label
Mo_occ
Mo_coef

For exemple :

BEGIN_PROVIDER [ character*(64), mo_label ]
  implicit none
  BEGIN_DOC
  ! Molecular orbital coefficients on AO basis set
  ! mo_coef(i,j) = coefficient of the ith ao on the jth mo
  ! mo_label : Label characterizing the MOS (local, canonical, natural, etc)
  END_DOC

  logical                        :: exists
  PROVIDE ezfio_filename
  call ezfio_has_mo_basis_mo_label(exists)
  if (exists) then
    call ezfio_get_mo_basis_mo_label(mo_label)
  else
    mo_label = 'no_label'
  endif
END_PROVIDER

And more qp_edit is broken. He didn't show the mos value.

Move SC2 outside of CISD

a.out

When you make build a a.out appear...

Create tests in MOs

Use Schwartz inequality for det selection

Function ``are_in_wavefunction``

Check if multiple determinants D are in the wf.

Sort D
For each D_i, binary search in the psi_det_sorted_bit array, when found remove from list
Alternate search of 1st element and last element of D to reduce the domain of search.

ocaml

make dans ocaml

rm -f -- test_atom
ocamlbuild -j 0 -syntax camlp4o -cflags "-g -warn-error A" -lflags "-g -warn-error A"  test_atom.byte  -use-ocamlfind  
+ ocamlfind ocamlc -c -g -warn-error A -thread -package core,sexplib.syntax,cryptokit -syntax camlp4o -o Point3d.cmi Point3d.mli
File "Point3d.mli", line 14, characters 26-50:
Error: Unbound module Qptypes.Positive_float
Command exited with code 2.
Compilation unsuccessful after building 12 targets (0 cached) in 00:00:00.
make[1]: *** [test_atom.byte] Error 10
make[1]: Leaving directory `/home/razoa/quantum_package/ocaml'
make: *** [build] Error 2

provide_mo_bielec_integral_jj: Error: Provider in an openMP section

`Pseudo` need `aux_basis` and `aux_basis` need `Pseudo`

If you check the EZFIO.cfg in Pseudo we need aux_basis.
And the module aux_basis need Pseudo.
Some cyclic dependency.

The idea to fix it is to allow aux_basis to write in the EZFIO folder pseudo. I will do this.

Density Matrix

Put 2DM in a map

quantum_package.rc:export:3: not valid in this context: PYTHONPATH+

I am using ZSH as my shell. When I try to source setup_environment.sh I get the following:

muammar@ihacku ⮀ ~/github/quantum_package ⮀ ⭠ master ⮀ . ~/github/quantum_package/setup_environment.sh
make: 'irpf90' is up to date.
quantum_package.rc:export:3: not valid in this context: PYTHONPATH+

If I print PYTHONPATH, it is empty in my environment. However, I paste here my sys.path:

muammar@ihacku ⮀ ~/github/quantum_package ⮀ ⭠ master ⮀ python -c "import sys; print sys.path"
['', '/usr/local/lib/python2.7/dist-packages/MeshPy-2011.1-py2.7-linux-x86_64.egg', '/usr/lib/python2.7', '/usr/lib/python2.7/plat-x86_64-linux-gnu', '/usr/lib/python2.7/lib-tk', '/usr/lib/python2.7/lib-old', '/usr/lib/python2.7/lib-dynload', '/usr/local/lib/python2.7/dist-packages', '/usr/local/lib/python2.7/dist-packages/setuptools-0.6c11-py2.7.egg-info', '/usr/lib/python2.7/dist-packages', '/usr/lib/python2.7/dist-packages/PILcompat', '/usr/lib/python2.7/dist-packages/gtk-2.0', '/usr/lib/pymodules/python2.7']

Wouldn't it be better to use sys.path.append instead?

Regards,

Optimize subroutine_diexcP

Replace 2**x by bit-shifts

'bld_ao_bielec_integrals_in_map_$THREAD_NUM' is being used without being defined

When compiling without openmp flag and launch SCF binary

AO map initialized
 Providing the AO integrals
forrtl: severe (193): Run-Time Check Failure. The variable 'bld_ao_bielec_integrals_in_map_$THREAD_NUM' is being used without being defined
Image              PC                Routine            Line        Source             
SCF                00000000005BE803  provide_ao_bielec         133  ao_bi_integrals.irp.F90
SCF                000000000049687A  bld_ao_bi_elec_in         616  Fock_matrix.irp.F90
SCF                000000000049B268  provide_fock_matr         519  Fock_matrix.irp.F90
SCF                000000000049E310  provide_fock_matr         734  Fock_matrix.irp.F90
SCF                000000000056D53F  huckel_guess_              41  huckel.irp.F90
SCF                000000000040377C  MAIN__                    107  SCF.irp.F90
SCF                00000000004032E6  Unknown               Unknown  Unknown
libc.so.6          00007F1A949C3EC5  Unknown               Unknown  Unknown
SCF                00000000004031D9  Unknown               Unknown  Unknown

EZFIO.cfg need more doc.

bug in selection ?

export OMP_NUM_THREADS=1
rm -r HBO.ezfio
source ~/quantum_package/quantum_package.rc
qp_create_ezfio_from_xyz -b "STO-3G" HBO.xyz
qp_edit -c HBO.ezfio
ezfio set_file HBO.ezfio
ezfio set integrals_bielec threshold_mo 1e-12
ezfio set integrals_bielec threshold_ao 1e-12
ezfio set perturbation do_pt2_end 0
ezfio set determinants n_det_max 10000
qp_run SCF HBO.ezfio
qp_set_mo_class HBO.ezfio  -core "[1-2]" -inact "[3-5]" -act "[6-9]" -virt "[10-11]"
qp_run cas_sd_selected HBO.ezfio
ezfio set determinants threshold_generators 1
ezfio set determinants read_wf 1
qp_run mrcc_cassd HBO.ezfio

avx and sse4.2 give different results

GFortran Support

For travis-cli, support of gfortran compiler need to be restored.

Precision in bielec_integrals_index_reverse

Precision in i2 = ceiling(0.5d0_(dsqrt(8.d0_dble(i1)+1.d0)-1.d0))
The result may be wrong. Check using bielec_integrals_index

lcpq / quantum_package Goto Github PK

quantum_package's People

Contributors

Stargazers

Watchers

Forkers

quantum_package's Issues

HOT FIX

Recommend Projects

Recommend Topics

Recommend Org