Fix: Add cuda arch 3.5 compilation support for ABACUS #2794
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@njzjz Please check if this PR solves your issue. Thanks! |
caic99
approved these changes
Aug 4, 2023
njzjz
approved these changes
Aug 4, 2023
hongriTianqi
approved these changes
Aug 4, 2023
denghuilu
added a commit
to denghuilu/abacus-develop
that referenced
this pull request
Aug 9, 2023
* Fix: error in specific situations refers to 2208 (deepmodeling#2742) * Test: simplify compiling ucell info in UNIT tests (deepmodeling#2747) * Paw : added accumulate_rhoij (deepmodeling#2734) * Added interface for calculation of rhoij * Added new test case * finished accumulate_rhoij, added test --------- Co-authored-by: wenfei-li <liwenfei@gmail.com> * Docs: fix dead links for vdw-d3 (deepmodeling#2741) fix deepmodeling#2711 * support reusing blacs_ctxt in Parallel_2D::set_desc (deepmodeling#2744) * Doc: add some details of UnitTest and Converge in Documentation (deepmodeling#2752) * add some details of UnitTest and Converge * Update docs/CONTRIBUTING.md Co-authored-by: Chun Cai <amoycaic@gmail.com> --------- Co-authored-by: Chun Cai <amoycaic@gmail.com> * Fix : update code for S(t+dt/2) (deepmodeling#2753) * update code for S(t+dt/2) and fix restart * update autotest * update autotest * Fix: change R in LCAO_Hamilt::calculate_HR_exx_sparse(), to alleviate the output Hexx(R) errors for few k-points (deepmodeling#2749) Co-authored-by: Mohan Chen <mohan.chen.chen.mohan@gmail.com> * Refactor: separate 2d-to-grid operation from `Local_Orbital_Charge` into an independent class `DMgamma_2dtoGrid` (deepmodeling#2748) * separate 2d-to-grid operation from Local_Orbital_Charge into an independent class * separate into new file --------- Co-authored-by: Mohan Chen <mohan.chen.chen.mohan@gmail.com> * Feature: Add a output interface for hcontainer and a sparse matrix class for CSR io (deepmodeling#2745) * Feature: Add CSR file reader classes in module_io (deepmodeling#2754) * refactor sparse_matrix class to use map * update output_hcontainer using new sparse matrix * add csr file reader * try to fix error * fix error --------- Co-authored-by: root <hongriTianqi> * Refactor(input model): create the input (deepmodeling#2671) * Refactor(input model): create the input rebuild class and function. This has no effect on the current code. Refactor deepmodeling#2549 * Refactor(input model): Fix the annotations in the last pull request. And fix a bug in unit test Refactor deepmodeling#2549 * Refactor(input model): create the input fix some compile bugs Refactor deepmodeling#2549 * fix namespace using bugs * fix some namespace using bugs * Update CMakeLists.txt fix conflict in io_print_info --------- Co-authored-by: Qianrui <76200646+Qianruipku@users.noreply.github.com> Co-authored-by: Xiaokuang Bai <baixiaokuang@gmail.com> * Fix : DeePKS for nspin = 2 and gamma point (deepmodeling#2755) * Fix : DeePKS for nspin = 2 and gamma point * Fixed cal_f_delta_gamma * fixed test --------- Co-authored-by: wenfei-li <liwenfei@gmail.com> Co-authored-by: dyzheng <zhengdy@dp.tech> * Fix issue deepmodeling#2721 (deepmodeling#2739) * Fix issue deepmodeling#2721: a correction on atom sequence check Fix issue deepmodeling#2721 * Update read_atoms.cpp * Update klist.cpp * Revert "Update klist.cpp" This reverts commit 0699982. --------- Co-authored-by: Qianrui <76200646+Qianruipku@users.noreply.github.com> Co-authored-by: Zhao Tianqi <hongriTianqi@users.noreply.github.com> * update version to v3.3.1 (deepmodeling#2766) * Update xc_functional_vxc.cpp (deepmodeling#2756) * Refactor: Two center integral implemented in module_nao (deepmodeling#2757) * add lookup function to TwoCenterTable * rabtab assumes uniform grid; table behavior change * two_center_integral initial commit * TwoCenterIntegrator implemented with test * Refactor: CubicSpline class stores derivatives instead of polynomial coefficients (deepmodeling#2772) * CubicSpline add support for derivative * function name fix * Refactor: add UnitTests and new functions and OverlapNew operator with HContainer (deepmodeling#2777) * update demo1_SR * Refactor: update demos in module_hcontainer * Test: add unit test for parallel_orbitals.cpp * update parallel_orbitals.h with get_indexes functions * Refactor: small modifies for constructor of OperatorLCAO * Refactor: update get_pointer() interface in AtomPair * Refactor: add func_folding and UnitTest * Feature: add new Operator for overlap of NAOs - OperatorNew * Fix: bug in CI * remove a useless head file in cal_dm.h (deepmodeling#2774) * Feature: add a parameter `symmetry_autoclose` to control the behavior when symmetry error occurs (deepmodeling#2740) * fix a debugging output in the 1st ion-step * add parameter symmetry_autoclose * Refactor: make `vl_grid_to_2D` and `Gint_inout` independent of `LCAO_Matrix` (deepmodeling#2758) * remove LCAO_Matrix dependence in vl_grid_to_2D * remove LCAO_Matrix dependence in Gint_inout --------- Co-authored-by: Zhao Tianqi <hongriTianqi@users.noreply.github.com> * Test: add a parallel script for parallel_2d_test (deepmodeling#2764) * Test: add a parallel script for parallel_2d_test * Update source/module_basis/module_ao/test/parallel_2d_test.sh --------- Co-authored-by: root <hongriTianqi> Co-authored-by: Peng Xingliang <91927439+pxlxingliang@users.noreply.github.com> * Fix: add a initialization of lmaxmax in UnitTests `cell_unitcell_test_setupcell` (deepmodeling#2767) * add a initialization of lmaxmax in UnitTests #test_setupcell * Update source/module_cell/test/unitcell_test_setupcell.cpp * Update source/module_cell/test_pw/unitcell_test_pw.cpp * Update source/module_cell/test/unitcell_test_setupcell.cpp * Update source/module_cell/test_pw/unitcell_test_pw.cpp --------- Co-authored-by: Zhao Tianqi <hongriTianqi@users.noreply.github.com> * Test: add parallel scripts for parallel_orbitals unit tests (deepmodeling#2778) Co-authored-by: root <hongriTianqi> * Test: add a unit test to constructor SR from reading csr file (deepmodeling#2780) Co-authored-by: root <hongriTianqi> * Docs: update conda-forge docs (deepmodeling#2784) * Fix: check atom types cpmsostence between STRU and DP model (deepmodeling#2770) * add cuda arch 3.5 compilation support for ABACUS (deepmodeling#2794) * Fix: bug in print_PAOs() (deepmodeling#2790) * Fix: autoset bx/by/bz when set nx/ny/nz manually (deepmodeling#2791) * Fix: bx = 0 if nx is set mannualy in INPUT * add a line of comment * add autoset bx/by/bz when set nx/ny/nz manually --------- Co-authored-by: root <hongriTianqi> * make RealGauntTable singleton & doxygen style change (deepmodeling#2789) Co-authored-by: Wenfei Li <38569667+wenfei-li@users.noreply.github.com> * new CubicSpline interface (deepmodeling#2800) Co-authored-by: Wenfei Li <38569667+wenfei-li@users.noreply.github.com> * Test: add a unit test for zhegv in lapack_connector (deepmodeling#2796) --------- Co-authored-by: dyzheng <zhengdy@dp.tech> Co-authored-by: Zhao Tianqi <hongriTianqi@users.noreply.github.com> Co-authored-by: Wenfei Li <38569667+wenfei-li@users.noreply.github.com> Co-authored-by: wenfei-li <liwenfei@gmail.com> Co-authored-by: Chun Cai <amoycaic@gmail.com> Co-authored-by: cyFortneu <33978601+maki49@users.noreply.github.com> Co-authored-by: wqzhou <33364058+WHUweiqingzhou@users.noreply.github.com> Co-authored-by: lyb9812 <88070104+lyb9812@users.noreply.github.com> Co-authored-by: PeizeLin <78645006+PeizeLin@users.noreply.github.com> Co-authored-by: Mohan Chen <mohan.chen.chen.mohan@gmail.com> Co-authored-by: GRYS <57103981+grysgreat@users.noreply.github.com> Co-authored-by: Qianrui <76200646+Qianruipku@users.noreply.github.com> Co-authored-by: Xiaokuang Bai <baixiaokuang@gmail.com> Co-authored-by: Yike Huang <67682086+kirk0830@users.noreply.github.com> Co-authored-by: jinzx10 <jzx016@hotmail.com> Co-authored-by: Omega <62006565+Satinelamp@users.noreply.github.com> Co-authored-by: Peng Xingliang <91927439+pxlxingliang@users.noreply.github.com> Co-authored-by: Yu Liu <77716030+YuLiu98@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
As mentioned by #2776, the pytorch flags would override the CMake CUDA_ARCHITECTURES, making the ABACUS build fail. This PR fix the compilation error with arch 3.5 by simply adding a custom atomicAdd function.
Close #2776.