A vote for new project：LibRPA

[Srlive1201]

[wanghan-iapcm]

+1

[wanghan-iapcm]

@Srlive1201 Is there a publicly available home page of the project? So the TOC member can better understand what is the project.

[likefallwind]

LibRPA

LibRPA is a standalone software aiming to calculate independent response function by space-time algorithm.

How to compile?

1. git clone or download the code package
2. modify 'make.inc' according your environment
3. make

Interfaced with FHI-aims

At this stage, LibRPA have interfaced with FHI-aims by writing and reading files:

• stru_out
• band_out
• KS_eigenvector_#.txt
• Cs_data_#.txt
• coulomb_mat_#.txt

The modified FHI-aims code files can be checked in FHI-aims-outifile-code/.
Consider the different versions of FHI-aims, you had better modify the FHI-aims files manual instead of recovering them.
Take care the new variable declarations in the beginning and the places of writing files. For convenience, we provide patch files in interfaces/FHI-aims. You may find the one appropriate
for your aims version.

Once you successfully finished FHI-aims calculation, the out-files will be ready for LibRPA.
Then run LibRPA in the same working directory.

Running LibRPA

LibRPA can be parallelized to hundreds of cores by MPI+openmp.
There are two input parameters needed to be care: 1) The minimax grids number; 2) The Green function threshold. When you run LibRPA, the two parameters needed to be offered, like:

$ mpirun /home/rongshi/LibRPA/chi0_main.exe 16 1e-4 > LibRPA_$workdir.$SLURM_JOB_ID.out

Other input parameters are parsed by specifying keyword and value pair in file librpa.in at the same directory. For example, if you want to set Coulomb threshold to 1e-5 and task to G0W0, you can write librpa.in as

task = g0w0
vq_threshold = 1e-5

If librpa.in and the related keyword is not found, the default value will be used.

Code design

[likefallwind]

I

LibRPA

LibRPA is a standalone software aiming to calculate independent response function by space-time algorithm.

How to compile?

1. git clone or download the code package
2. modify 'make.inc' according your environment
3. make

Interfaced with FHI-aims

At this stage, LibRPA have interfaced with FHI-aims by writing and reading files:

• stru_out
• band_out
• KS_eigenvector_#.txt
• Cs_data_#.txt
• coulomb_mat_#.txt

The modified FHI-aims code files can be checked in FHI-aims-outifile-code/. Consider the different versions of FHI-aims, you had better modify the FHI-aims files manual instead of recovering them. Take care the new variable declarations in the beginning and the places of writing files. For convenience, we provide patch files in interfaces/FHI-aims. You may find the one appropriate for your aims version.

Once you successfully finished FHI-aims calculation, the out-files will be ready for LibRPA. Then run LibRPA in the same working directory.

Running LibRPA

LibRPA can be parallelized to hundreds of cores by MPI+openmp. There are two input parameters needed to be care: 1) The minimax grids number; 2) The Green function threshold. When you run LibRPA, the two parameters needed to be offered, like:

$ mpirun /home/rongshi/LibRPA/chi0_main.exe 16 1e-4 > LibRPA_$workdir.$SLURM_JOB_ID.out

Other input parameters are parsed by specifying keyword and value pair in file librpa.in at the same directory. For example, if you want to set Coulomb threshold to 1e-5 and task to G0W0, you can write librpa.in as

task = g0w0
vq_threshold = 1e-5

If librpa.in and the related keyword is not found, the default value will be used.

Code design

Just copy the README files here. And the repo is: https://github.com/Srlive1201/LibRPA

[KuangYu]

+1

[Srlive1201]

@Srlive1201 Is there a publicly available home page of the project? So the TOC member can better understand what is the project.

Sorry for our home page is still very rudimentary. We will improve it in the near future. For more details, please refer to the documentations or slides in \docs.

[mohanchen]

+1

[likefallwind]

LibRPA team want to do some work before entering into deepmodeling, so this vote stops until the team is ready entering deepmodeling.

[jameswind]

+1 Regarding the homepage, once the project is approved by deepmodeling, we may pay more attention to the readme and doc to improve it.

On Thu, Sep 1, 2022 at 10:38 PM Rong Shi @.***> wrote:

@Srlive1201 https://github.com/Srlive1201 Is there a publicly available home page of the project? So the TOC member can better understand what is the project.

Sorry for our home page is still very rudimentary. We will improve it in the near future. For more details, please refer to the documentations or slides in \docs https://github.com/Srlive1201/LibRPA/tree/master/docs.

— Reply to this email directly, view it on GitHub https://github.com/deepmodeling/community/pull/16#issuecomment-1234374286, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEJ6DC6VXJVEBWMSLKSBA33V4C5WDANCNFSM6AAAAAAQA7HDHA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[jameswind]

+1