Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Tech isolv bse #141

Open
wants to merge 105 commits into
base: tech-master
Choose a base branch
from
Open

Tech isolv bse #141

wants to merge 105 commits into from

Conversation

sangallidavide
Copy link
Member

This branch contains the latest development related to the Isolv-BSE project.

It has been kept on hold, waiting for the new release, e.g. version 3.22, of the slepc library (partner of the Isolv-bse project.
No the the library is available, this can also enter in the main release process of the code

sangallidavide and others added 30 commits June 1, 2024 22:29
MODIFIED *  configure include/version/version.m4 lib/archive/package.list

Changes:
- [yambo] Slepc and Petsc version updated to 3.21

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 bse/K_blocks.F bse/K_kernel.F interface/INIT_load.F io_parallel/io_BS_PAR_block.F io_parallel/io_BS_PAR_init.F modules/mod_BS.F

Changes:
- [yambo] For runs where the BSE matrix is only constructed, e.g. no
  solver, the BS_blk(iB)%mat are allocated and de-allocated withing the loops
  over blocks. This saves a lot of memory if used together with the new
  I/O mode BSKIOmode="write_all_steps"

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 bse/K_kernel.F bse/K_solvers.F

Changes:
- [yambo] New mode, controlled with "write_all_steps" is activated also
  in presence of solvers. In such case, during the calculation of the
  kernel the blocks are not kept in memory, and only later, when
  entering the solver, the whole kernel is loaded into memory

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 modules/mod_BS.F

Changes:
- Subroutine for cleaning all blocks uses newly created subruntine which
  cleans single block

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4

Changes:
- [yambo]: configure re-generated after merge

Patch sent by:  Davide Sangalli <[email protected]>
This new type is a "nest" matrix for BSE with two shell blocks R and C.
Requires Mult operations for the shell blocks.
MODIFIED *  include/version/version.m4 bse/K_driver.F bse/K_driver_init.F bse/K_multiply_by_V.F bse/K_shell_matrix.F

Changes:
- [yambo] Changing K_multiply_by_V structure

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 bse/.objects bse/K_Haydock.F bse/K_diago_driver.F bse/K_driver_init.F bse/K_multiply_by_V.F bse/K_multiply_by_V_slepc.F bse/K_multiply_by_V_transpose_slepc.F bse/K_shell_matrix.F

NEW *  bse/K_multiply_by_V_slepc_nest.F

Changes:
- [yambo] First working version of nested slepc solver with shell approach

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 lib/slepc/Makefile.loc bse/K_diago_driver.F bse/K_stored_in_a_nest_matrix.F linear_algebra/MATRIX_slepc.F

NEW *

Changes:
- [yambo] Commented call to Slepc get bse type, otherwise it always
  overwirets what in input

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 bse/K_stored_in_a_nest_matrix.F linear_algebra/MATRIX_slepc.F

NEW *

Changes:
- [yambo] Few changes in view of shift and invert approach to slepc

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 bse/K_blocks.F bse/K_kernel.F bse/K_solvers.F bse/K_stored_in_a_big_matrix.F bse/K_stored_in_a_nest_matrix.F bse/K_stored_in_a_slepc_matrix.F modules/mod_BS.F

Bugs:
- [ypp] compilation fixed

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 linear_algebra/MATRIX_slepc.F

Changes:
- [yambo] Added missing set bse type

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 bse/K_diago_driver.F bse/K_stored_in_a_nest_matrix.F bse/K_stored_in_a_slepc_matrix.F

NEW *

Changes:
- [yambo] Added timing to K_stored_in_a_slepc/nest_matrix
- [yambo] After slepc matrix redistribution, BSE kernel is de-allocated
- [yambo] Improved messaging for shell vs explicit algorithm

Patch sent by:  Davide Sangalli <[email protected]>
sangallidavide and others added 23 commits December 20, 2024 22:39
MODIFIED *  configure include/version/version.m4 lib/qe_pseudo/DOUBLE_project.dep lib/slatec/DOUBLE_project.dep

Changes:
- fixed .dep files

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 pol_function/X_redux.F

Bugs:
 -Fixed a small bug in parallelization with intel

Additions:
-

Changes:
-

Patch sent by:  Claudio Attaccalite <[email protected]>
MODIFIED *  configure include/version/version.m4

Additions:
 -Version updated

Changes:
-

Patch sent by:  Claudio Attaccalite <[email protected]>
MODIFIED *  configure include/version/version.m4 pol_function/X_redux.F

Bugs:
 -Fixed parallel bug with intel

Additions:
-

Changes:
-

Patch sent by:  Claudio Attaccalite <[email protected]>
MODIFIED *  configure include/version/version.m4 pol_function/X_redux.F

Bugs:
 -Sorry I did a commit in the wrong branch, reverted

Additions:
-

Changes:
-

Patch sent by:  Claudio Attaccalite <[email protected]>
MODIFIED *  configure include/version/version.m4 interpolate/INTERPOLATION_BZ_coeff.F

Bugs:
 -Fixed another bug in parallelization but there is still a failure with
  ifor

Additions:
-

Changes:
-

Patch sent by:  Claudio Attaccalite <[email protected]>
MODIFIED *  config/m4/netcdf_f90.m4 configure include/version/version.m4

Changes:
- [yambo] Tentative fix for issue detected in
  #168 with netcdf detection

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 collisions/COLLISIONS_load.F interface/INIT_load.F io_parallel/io_COLLISIONS_header.F nloptics/NL_initialize.F

Bugs:
 -Fixed collisions cutoff in NL, it was always zero and not possible to
  change it

Additions:
-

Changes:
-

Patch sent by:  Claudio Attaccalite <[email protected]>
MODIFIED *  configure include/version/version.m4 collisions/COLLISIONS_load.F

Bugs:
 -Small error in output

Additions:
-

Changes:
-

Patch sent by:  Claudio Attaccalite <[email protected]>
MODIFIED *  configure include/version/version.m4 collisions/COLLISIONS_load.F

Bugs:
 -Restored old output

Additions:
-

Changes:
-

Patch sent by:  Claudio Attaccalite <[email protected]>
Make the links in README absolute to fix their rendering
MODIFIED *  configure include/version/version.m4 nloptics/el_density_vbands.F

Bugs:
 -Fixed compilation with ifort

Additions:
-

Changes:
-

Patch sent by:  Claudio Attaccalite <[email protected]>
MODIFIED *  configure include/version/version.m4

Changes:
- condigure re-generated after merge

Patch sent by:  Davide Sangalli <[email protected]>
MODIFIED *  configure include/version/version.m4 modules/mod_openmp.F

Changes:
- openblas max number of threads commented as in master
  It would need an #if defined _OPENBLAS ...

Patch sent by:  Davide Sangalli <[email protected]>
Copy link
Contributor

@attacc attacc left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fine for me, are new tests available for this fork?

@sangallidavide sangallidavide requested a review from myrta January 20, 2025 13:34
@sangallidavide
Copy link
Member Author

Yeah, there is a connected test-suite branch with few new input files for BSE with coupling which can be found here:
https://github.com/yambo-code/yambo-tests/tree/tech-isolve-bse/TESTS/MAIN/hBN/GW-OPTICS/INPUTS

Including the new "Gruning" solver coded by the slepc team :-)
https://github.com/yambo-code/yambo-tests/blob/tech-isolve-bse/TESTS/MAIN/hBN/GW-OPTICS/INPUTS/04_bse_slepc_cpl_expl_Gruning

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants