Skip to content

Cherry-pick #6337 to resolve the issue where the LTS branch doesn't compile on CUDA Version >= 12.9 (issue #6443). #6584

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.

Sign up for GitHub

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

Jump to bottom
Open
wants to merge 1 commit into
base: LTS
Choose a base branch
from
Open

Cherry-pick #6337 to resolve the issue where the LTS branch doesn't compile on CUDA Version >= 12.9 (issue #6443). #6584

wants to merge 1 commit into from

Conversation

zgn-26714
Copy link

fix compile error on CUDA >= 12.9

@Flying-dragon-boxing @zhangzh-pku
@WHUweiqingzhou @kirk0830 @haozhihan @hongriTianqi @PeizeLin @jinzx10 @caic99 @pxlxingliang @jieli-matrix @wenfei-li @denghuilu @yizeyi18 @jingan-181 @grysgreat @mohanchen
Fix: Fix Compile Error on CUDA Version > 12.9 (deepmodeling#6337)
60e36b3 
* feat pexsi

* fix : diag not completed

* feat

* feat: pexsi hsolver

* CMake building implemented

* Works

* adapt to the new container

* Turn off USE_PEXSI

* Update LibRI to 553c91c

* modify include files

* namespace-ize

* new inputs added

* Configure Makefile Compiling, fix typos

* Fix Makefile Intel toolchains compile errors

* Fix even more PEXSI related Makefile compiling issues

* Modify inputs and update to latest version (#2)

* run INPUT.Default() in every process in InputParaTest (deepmodeling#3490)

Co-authored-by: kirk0830 <[email protected]>

* add blas support for FindLAPACK.cmake (deepmodeling#3497)

* more unittest of QO: towards orbital selection (deepmodeling#3499)

* Fix: fix bug in mulliken charge calculation (deepmodeling#3503)

* fix phase

* fix case test

* Refactor: namespace Conv_Coulomb_Pot_K (deepmodeling#3446)

* Refactor: namespace Conv_Coulomb_Pot_K

* Refactor: namespace Conv_Coulomb_Pot_K

---------

Co-authored-by: wqzhou <[email protected]>

* enable the computation of all zeros in one function call (deepmodeling#3449)

Co-authored-by: wqzhou <[email protected]>

* replace ios.eof() by ios.good() to avoid meeting badbit and failbit in reading STRU (deepmodeling#3506)

* Build: add ccache to accelerate the testing process (deepmodeling#3509)

* Build: add ccache to accelerate the testing process

* Update test.yml

* Update test.yml

* Update test.yml

* Docs: to avoid the misunderstanding in docs (deepmodeling#3518)

* to avoid the misunderstanding in docs

* Update docs/quick_start/hands_on.md

Co-authored-by: Chun Cai <[email protected]>

---------

Co-authored-by: Chun Cai <[email protected]>

* Docs: fix a missing depencency in conda build env (deepmodeling#3508)

* Feature: Add ENABLE_RAPIDJSON option to control the output of abacus.json (deepmodeling#3519)

Add ENABLE_RAPIDJSON option to control the output of abacus.json

* Feature: add python wrapper for math sphbes (deepmodeling#3475)

* recommit for review

* add python wrapper

* remove timer since performace tests add

* Feature: support segment split in kline mode in KPT file and `out_band` band output precision control, `8` as default (deepmodeling#3493)

* add precision control

* correct serial version of nscf_band function

* fix issue 3482

* update unit and integrated test

* update document

* correct unittest and make compatible with false and true

* fix: bug in Autotest.sh when result.ref has no totaltimeref (deepmodeling#3523)

* Fix : unit test of module_xc (deepmodeling#3524)

* Fix: omit small magnetic moments to avoid numerical instability (deepmodeling#3530)

* update deltalambda

* avoid numerical error in orbMulP

* add constrain on Mi

* change case reference value

* Fix: fix multiple compiler warnings (deepmodeling#3515)

* Fix: add noreturn attribute to warning_quit

* Add type conversion

* fix string literal

* fix small number trunctuation

* Fix system call returned value not checked

* fix missing braket

* Refactor parameter_pool.cpp and parameter_pool.h

* remove duplicated return statements

* Change WARNING_QUIT occurances in tests

* Add warning message to help debug UT

* output the default precision flag (deepmodeling#3496)

Co-authored-by: kirk0830 <[email protected]>

* Build: Improving CMake performance for finding LibXC and ELPA (deepmodeling#3478)

* Fix for finding LibXC and ELPA

* For compatibility to previous routines

* syntax fix for FindELPA.cmake

* Update cmake/FindELPA.cmake

Co-authored-by: Chun Cai <[email protected]>

* Using CMake interface as default for finding LibXC

* update docs

* fix for FindLibxc: changing imcompatible if statement

* fix for FindLibxc: changing imcompatible if statement

* fix for FindLibxc: changing imcompatible if statement

* update docs for installing pkg-config

* Update FindLibxc.cmake

* Update FindLibxc.cmake

* remove previous LibXC routine in CMakeLists.txt

Co-authored-by: Chun Cai <[email protected]>

* Update easy_install.md with Makefile-built LibXC supported

* Update easy_install.md to include different behavior in different version on finding ELPA

---------

Co-authored-by: Chun Cai <[email protected]>

* Docs: correct some docs about mp2 smearing method (deepmodeling#3533)

* correct some docs about mp2 smearing method

* add docs about mv method

* Feature : printing band density (deepmodeling#3501)

Co-authored-by: wenfei-li <[email protected]>
Co-authored-by: wqzhou <[email protected]>

* add some docs for PR#3501 (deepmodeling#3537)

* Feature: enable restart charge density mixing during SCF (deepmodeling#3542)

* add a new parameter mixing_restart

* do not update rho if iter==mixing_restart

* do not update rho if iter==mixing_restart-1

* reset mix and rho_mdata if iter==mixing_restart

* fix SCF exit directly since drho=0 if iter=GlobalV::MIXING_RESTART

* re-set_mixing in eachiterinit for PW and LCAO

* enable SCF restarts in esolver_ks::RUN

* add some UnitTests

* add some Docs

* new inputs added

* Update input-main.md (deepmodeling#3551)

Solve the format problem mentioned in issue 3543

* Build: fix compatibility issue against toolchain install (deepmodeling#3540)

* Fix for finding LibXC and ELPA

* For compatibility to previous routines

* syntax fix for FindELPA.cmake

* Update cmake/FindELPA.cmake

Co-authored-by: Chun Cai <[email protected]>

* Using CMake interface as default for finding LibXC

* update docs

* fix for FindLibxc: changing imcompatible if statement

* fix for FindLibxc: changing imcompatible if statement

* fix for FindLibxc: changing imcompatible if statement

* update docs for installing pkg-config

* Update FindLibxc.cmake

* Update FindLibxc.cmake

* remove previous LibXC routine in CMakeLists.txt

Co-authored-by: Chun Cai <[email protected]>

* Update easy_install.md with Makefile-built LibXC supported

* Update easy_install.md to include different behavior in different version on finding ELPA

* fix compatibility issue against toolchain

* Change default ELPA install routine to old one

---------

Co-authored-by: Chun Cai <[email protected]>

* Test: Configure performance tests for math libraries (deepmodeling#3511)

* add performace test of sphbes functions.

* fix benchmark cmake errors

* add dependencies for docker

* update docs

* add performance tests for sphbes

* add google benchmark

* rewrite benchmark tests in fixtures

* disable internal testing in benchmark

* merge benchmark into integration test

---------

Co-authored-by: StarGrys <[email protected]>

* Configure Makefile Compiling, fix typos

* Fix Makefile Intel toolchains compile errors

* Fix even more PEXSI related Makefile compiling issues

* Update hsolver_pw.cpp (deepmodeling#3556)

when use_uspp==false, overlap matrix should be E.

* Fix: cuda build target (deepmodeling#3276)

* Fix: cuda buid target

* Update CMakeLists.txt

---------

Co-authored-by: Denghui Lu <[email protected]>

---------

Co-authored-by: wqzhou <[email protected]>
Co-authored-by: kirk0830 <[email protected]>
Co-authored-by: Haozhi Han <[email protected]>
Co-authored-by: Zhao Tianqi <[email protected]>
Co-authored-by: PeizeLin <[email protected]>
Co-authored-by: jinzx10 <[email protected]>
Co-authored-by: Chun Cai <[email protected]>
Co-authored-by: Peng Xingliang <[email protected]>
Co-authored-by: Jie Li <[email protected]>
Co-authored-by: Wenfei Li <[email protected]>
Co-authored-by: Denghui Lu <[email protected]>
Co-authored-by: YI Zeping <[email protected]>
Co-authored-by: wenfei-li <[email protected]>
Co-authored-by: jingan-181 <[email protected]>
Co-authored-by: StarGrys <[email protected]>
Co-authored-by: Haozhi Han <[email protected]>

* Revert "Modify inputs and update to latest version"

* Update FindPEXSI.cmake to fix Comments

* Fix CI errors

* Fix CI Errors and Merge with Upstream

* Resolve Pull Request Reviews

* Fix parallel communication related issue

* Fix vars in Makefile.vars, add input tests and comments for pexsi vars

* Fix nspin > 1 cases

* Improvement: take calculated mu as new initial guess, may slightly improve performance

* Fix mistakes in the last commit

* Fix: params and features
- set default pexsi_temp
- fix md in pexsi

* fix empty lines

* Fix: move params to pexsi_solver, rename USE_PEXSI to ENABLE_PEXSI

* Docs: added docs for pexsi inputs

* Fix unit test issues in input_conv

* Change default pexsi_npole from 80 to 40

* Place pexsi_EDM in DensityMatrix, set size of pexsi_dm = 1 when GlobalV::NSPIN==4, and add comments for dmToRho

* An unit test added for DiagoPexsi

* modify for changed gint interface

* correct nspin related behaviors

* add efermi passthrough

* Revert "add efermi passthrough"

This reverts commit d7b402d.

* commits to resolve conversations related to codes

* DM and EDM pointers in pexsi now handled by diagopexsi, and copying h s matrices no longer needed

* add pexsi examples

* fix pexsi unit test (original version shouldn't run)

* add building docs for pexsi

* set cxx standard to c++14, which is required in make_unique

* Fix: Fix typo related to pexsi

* update to PPEXSIDFTDriver2

* default npoints to 1, so single core pexsi will work

* Fix Compile errors

* refactor to abandon `pdiagh`

* Fix mu_buffer and nspin

* Updates with latest

* Refactor: in ESolver_KS_PW, calculate deband in iter_finish, not in hamilt2density

* Fix: make files in consistent with upstream

* Refactor

* Refactor

* Refactor

* Refactor

* Refactor

* Refactor: fix unit test

* Refactor: fix unit test

* Refactor: fix unit test

* Refactor: fix unit test

* Refactor: Remove set kvec funcs in `K_Vectors`

* Refactor: Remove final_scf

* Refactor: Fix kvecc2d/d2c

* Fix: Tests

* Fix: Tests

* Fix: Tests

* Fix: Tests

* Refactor: Final?

* Fix

* Fix

* Fix

* Fix

* Fix: Compile Error on CUDA > 12.9

* Fix: Compile Error on CUDA > 12.9

---------

Co-authored-by: zhangzhihao <[email protected]>
Co-authored-by: zhangzh-pku <[email protected]>
Co-authored-by: wqzhou <[email protected]>
Co-authored-by: kirk0830 <[email protected]>
Co-authored-by: Haozhi Han <[email protected]>
Co-authored-by: Zhao Tianqi <[email protected]>
Co-authored-by: PeizeLin <[email protected]>
Co-authored-by: jinzx10 <[email protected]>
Co-authored-by: Chun Cai <[email protected]>
Co-authored-by: Peng Xingliang <[email protected]>
Co-authored-by: Jie Li <[email protected]>
Co-authored-by: Wenfei Li <[email protected]>
Co-authored-by: Denghui Lu <[email protected]>
Co-authored-by: YI Zeping <[email protected]>
Co-authored-by: wenfei-li <[email protected]>
Co-authored-by: jingan-181 <[email protected]>
Co-authored-by: StarGrys <[email protected]>
Co-authored-by: Haozhi Han <[email protected]>
Co-authored-by: Mohan Chen <[email protected]>
@mohanchen mohanchen added the Long-Time Support (LTS) Issues related to LTS version label Oct 11, 2025
@zgn-26714 zgn-26714 changed the title Cherry-pick #6392 to resolve the issue where the LTS branch doesn't compile on CUDA Version >= 12.9 (issue #6443). Cherry-pick #6337 to resolve the issue where the LTS branch doesn't compile on CUDA Version >= 12.9 (issue #6443). Oct 12, 2025
@zgn-26714
Copy link
Author

The test for case 204_NO_KP_NC passes locally.
image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

Long-Time Support (LTS) Issues related to LTS version

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@zgn-26714 @mohanchen @Flying-dragon-boxing