Commit Graph

978 Commits

Author SHA1 Message Date
d9abc3fcc0 update CUDA Toolkit / GPU compatibility lists and GPU package compilation settings 2023-01-03 11:56:44 -05:00
e033cebcdd update lammps-user-pace version and checksum 2023-01-03 14:30:04 +01:00
8610fc6d33 simplify intel versions of electrode fixes 2023-01-03 12:32:12 +10:00
c9a3894d12 grammar 2022-12-28 17:11:18 -05:00
d47acfc0c4 Have PyTorch interface for MLIAP working in Kokkos. This uses cuPy and a simple example is provided 2022-12-28 07:01:47 +01:00
afae6222b0 Update build instructions for COLVARS package 2022-12-27 20:02:44 -05:00
989ec1b859 remove lmp/LMP_ prefix from Lepton namespace and files to share it with colvars 2022-12-27 10:57:43 -05:00
cebb97e790 Merge branch 'develop' after the distributed-grids PR was merged into amoeba-gpu, noted some API changes in reverse_comm 2022-12-24 00:41:07 -06:00
adb27c6e3c Update CMake 2022-12-23 08:56:00 -07:00
749adf3a59 one more tweak to allow more x86 platforms to use JIT with Lepton 2022-12-23 05:32:35 -05:00
acf683e9d0 define ASMJIT_STATIC to work around Windows issues 2022-12-23 01:35:05 -05:00
3a6492fc42 use JIT compiler only on Linux for now 2022-12-22 23:46:21 -05:00
a8c881aaf3 try to address linker issues with asmjit on older Linux machines 2022-12-22 23:09:09 -05:00
e59f99b440 add support for JIT compilation 2022-12-22 22:50:01 -05:00
91c498c413 suppres explicit exports/import in Lepton lib 2022-12-22 16:32:17 -05:00
5f934e3eae add LEPTON package build system support for CMake 2022-12-21 18:28:35 -05:00
281e67d6fb Merge branch 'develop' into distributed-grids 2022-12-14 18:12:10 -05:00
8af77c690c Merge branch 'develop' into amoeba-gpu 2022-12-14 13:16:41 -06:00
88ac09a8c0 build system support: fix sgcmc may only be compiled if EAM is available 2022-12-13 16:35:19 -05:00
6365af8704 Merge branch 'develop' into distributed-grids 2022-12-01 10:38:55 -05:00
6c3ebc19b7 Merge branch 'develop' into gran-temp 2022-11-30 08:54:19 -05:00
b2c5c1fda4 Fixing cmake style header 2022-11-10 21:06:47 -07:00
d1dd8d3c11 Renaming files/classes, other small, misc updates 2022-11-10 16:06:36 -07:00
01b4600ba5 remove obsolete file 2022-11-10 07:11:22 -05:00
710197cd88 add MS-MPI support to CMake support for plugins 2022-11-04 23:00:34 -04:00
0de50f29f7 add option to use the MS-MPI SDK to cross-compile Windows binaries 2022-11-04 21:56:06 -04:00
b4118c51cc merged in current master 2022-11-04 08:22:18 -06:00
74122040a3 add cmake support for generating gsm style header file 2022-11-02 20:20:06 -04:00
10d3b1a875 Merge pull request #3510 from akohlmey/collected-small-changes
Collected small changes for the next patch release
2022-11-02 19:52:06 -04:00
902a07a83a update for portability 2022-11-01 21:23:45 -04:00
3252424dea Merge branch 'develop' of https://github.com/lammps/lammps into kokkos-mliap-so3 2022-10-31 15:20:38 -07:00
6afa1d14d2 downgrade KOKKOS OpenMP check to version 3.1
need to apply special exception for NVHPC/PGI compilers
2022-10-26 22:36:15 -04:00
ca6aba52bb modernize OpenMP detection and check for omp.h in CMake 2022-10-26 22:35:41 -04:00
511e1bb7b9 Merge branch 'develop' into kokkos-mliap-so3 2022-10-24 19:03:42 -04:00
d313179917 add compatibility to VTK version 9.0 and above 2022-10-24 16:25:10 -04:00
005d72bf76 Merge branch 'develop' into kokkos-mliap-so3 2022-10-24 04:26:37 -04:00
2c4aa42c31 conditional compilation of ML-IAP package KOKKOS components with CMake 2022-10-24 04:22:48 -04:00
2aa1dfa6d6 replace one more suffix 2022-10-22 22:28:12 -04:00
caeaeb5d33 Python support in ML-IAP requires NumPy. Check for it if CMake supports it. 2022-10-21 15:31:41 -04:00
5c4bb611d1 find cythonize executable on recent FreeBSD versions 2022-10-21 11:25:13 -04:00
4d2661433f try to make more portable (in case this ever gets ported to windows) 2022-10-21 09:28:46 -04:00
ef83f78e1e add md5sums for plumed 2.7.5 and 2.8.1, update default version to 2.8.1 2022-10-21 09:28:15 -04:00
ce50300534 Created a Kokkos version of the ML-IAP routines. Completed one model and
one descriptor along with the pair style and data member.  To test the stack
one can run

lmp -k on g 1 -sf kk -pk kokkos neigh full newton on -in in.mliap.so3.Ni_Mo -echo both

and compare it to the non-kokkos results.
Results are identical for cuda/openmp/serial and non-kokkos
2022-10-21 09:45:18 +02:00
c10b5b02c0 Merge branch 'develop' into collected-small-changes 2022-10-17 22:40:01 -04:00
2f60db4aba Merge remote-tracking branch 'upstream/develop' into upd/refactor-includes
# Conflicts:
#	cmake/Modules/Packages/ML-PACE.cmake
2022-10-15 18:07:07 +02:00
33681b5d84 WIP: update ML-PACE.cmake 2022-10-15 18:06:07 +02:00
a218c2ad9c Fix CMake build 2022-10-12 11:33:55 -06:00
c741eca953 detect clang-format version on Ubuntu 22.04LTS 2022-10-12 12:31:33 -04:00
c113253e2d Merge branch 'develop' of https://github.com/lammps/lammps into kk_update_3.7 2022-10-10 13:44:02 -07:00
1fb07387b9 Merge pull request #3479 from yury-lysogorskiy/feature/pace-al
BUGFIX: address issue with  compilation KOKKOS + pace/extrapolation
2022-10-10 14:57:13 -04:00