Skip to content

Implement CVC algorithm and enhance LR class#11

Open
Fisherd99 wants to merge 5 commits into
abacusmodeling:masterfrom
Fisherd99:master
Open

Implement CVC algorithm and enhance LR class#11
Fisherd99 wants to merge 5 commits into
abacusmodeling:masterfrom
Fisherd99:master

Conversation

@Fisherd99

Copy link
Copy Markdown

1.add cvc_mo algorithm

  • construct C_mo on the fly
  • construct A and B matrix for BSE/LR

2.add physics class Hartree

3.add cell_nearest_direction fucntion

  • brute-force search nearest cell
  • handle the same nearest cell with direction judge

4.modify LR.h

  • add members of k/q-point to support non-uniform BZ sampling
  • temporarily cancel comm_map2 in LR.h since comm fails for large tensor

Unit test is provided for point 3 and 4.

maki49 added 4 commits July 20, 2026 05:01
commit f6a2205
Author: maki49 <1579492865@qq.com>
Date:   Sun Sep 28 10:48:14 2025 +0800

    construct LR from Exx

commit 7a960fd
Author: maki49 <1579492865@qq.com>
Date:   Sun Sep 21 14:41:38 2025 +0800

    add physics/LR
commit bd40883
Author: maki49 <1579492865@qq.com>
Date:   Tue Sep 30 17:49:30 2025 +0800

    remove timer and title

commit ae7a76d
Author: maki49 <1579492865@qq.com>
Date:   Fri Sep 26 21:15:49 2025 +0800

    minimize the index access

commit cb68103
Author: maki49 <1579492865@qq.com>
Date:   Fri Sep 26 21:03:58 2025 +0800

    wait after each term to save memory

commit 4793d3c
Author: maki49 <1579492865@qq.com>
Date:   Wed Sep 24 13:57:26 2025 +0800

    malloc_trim(0)

commit 150b929
Author: maki49 <1579492865@qq.com>
Date:   Wed Sep 24 13:43:32 2025 +0800

    fix: filter zero-tensors

commit 3a773e4
Author: maki49 <1579492865@qq.com>
Date:   Tue Sep 23 21:27:58 2025 +0800

    symmetry filter

commit 811e6da
Author: maki49 <1579492865@qq.com>
Date:   Tue Sep 23 21:27:22 2025 +0800

    default cvc (only for test)

commit 3f4b6b9
Author: maki49 <1579492865@qq.com>
Date:   Tue Sep 23 19:02:13 2025 +0800

    omp parallel method 3

commit d91c80d
Author: maki49 <1579492865@qq.com>
Date:   Tue Sep 23 17:31:38 2025 +0800

    omp parallel method 2

commit 0a8a529
Author: maki49 <1579492865@qq.com>
Date:   Tue Sep 23 16:08:48 2025 +0800

    fix omp bug

commit e329128
Author: maki49 <1579492865@qq.com>
Date:   Sat Sep 20 16:58:40 2025 +0800

    CVC (D-perspective) algorithm (draft)

    fix adding tensor maps

    fix compile problems

commit d4f7320
Author: Minye Zhang <minyez.physchem@gmail.com>
Date:   Mon May 19 17:25:38 2025 +0200

    remove redundant member variables
@maki49

maki49 commented Jul 20, 2026

Copy link
Copy Markdown
Contributor

Large irrelevant difference in ri/LRI.h and unittests/ri/Cell_Nearest-test.hpp. Maybe unix2dos or dos2unix can help.

- construct C_mo on the fly to reduce memory consumption
- rearrange order of multiply to avoid permute
- use lapack in Csk_ao_mo_to_Cs_mo

2.add Hartree

3.add cell_nearest_direction
- brute-force search nearest cell
- handle the same nearest cell with direction judge
4.modify LR.h
- temporarily cancel comm_map2 in LR.h since comm fails for large tensor
@Fisherd99

Copy link
Copy Markdown
Author

Large irrelevant difference in ri/LRI.h and unittests/ri/Cell_Nearest-test.hpp. Maybe unix2dos or dos2unix can help.

Thanks for pointing out. I fixed it with unix2dos, transforming LF to CRLF for line-endings.

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.

2 participants