Commit Graph

356 Commits

Author SHA1 Message Date
a9b794b2ab Updates to the core-shell gpu pair styles 2018-06-20 10:43:13 -05:00
3ae8fdccd8 Updated the kernels for born/coul/long/cs and coul/long/cs gpu styles 2018-06-19 16:22:00 -05:00
70a7b37614 Updated pair coul/long/cs and born/coul/long/cs; updated gpu neighbor builds to support core-shell styles where r2 can be tiny. 2018-06-19 15:50:02 -05:00
6842a527e0 Fixed bugs and added born/coul/long/cs/gpu and coul/long/cs/gpu, need tests 2018-06-18 00:15:28 -05:00
265b6c261e Fixed bugs with lj/expand/coul/long and its gpu version 2018-06-15 23:38:37 -05:00
a4a3133b76 Switched to standard C++ header files for consistency with other source files 2018-06-01 23:13:49 -05:00
341fa160fe Updated UCL_Device built with OpenCL to use platforms that support accelerators by default 2018-05-23 16:11:55 -05:00
2f9e6d4566 Fixed bugs in lal_device.cpp with node_names dynamically allocated and dipole/long/gpu 2018-05-21 23:32:25 -05:00
838a1938bb Added pair styles lj/expand/coul/long to USER-MISC and its gpu version to GPU package, added lj/cut/dipole/long/gpu; and added an example Makfile to build GPU package for multi-arch 2018-05-20 17:03:12 -05:00
fccc16f42c tune OpenCL for generic architecture by default 2018-05-08 00:50:23 -04:00
e7e1827e57 remove variable length arrays in gpu lib 2018-05-08 00:48:53 -04:00
3394d18a2a convert gpu lib to use c++-style include files 2018-05-08 00:25:10 -04:00
776f588746 add support for UFM pair style support to OpenCL compilation makefile 2018-02-12 18:43:30 +01:00
4af14becb5 Merge pull request #710 from rpleite/add-pair-ufm
Add the UFM pair potential - (GPU/OPT/OMP)
2018-02-02 14:38:28 -07:00
b34000a5e1 make *_gpu_reinit() functions consistent in returning void. correct prototypes in src/GPU, too.
this supersedes and closes #719
2017-11-04 02:49:22 -04:00
941ee565a1 Added lal_ufm_ext.cpp in lib/gpu 2017-10-24 11:12:51 -02:00
b63acf6843 Added the UFM files (doc/src - lib/gpu - src) 2017-10-24 11:11:10 -02:00
be96486d66 compile all objects in the GPU library with -fPIC for building shared object 2017-09-19 11:24:48 -04:00
48dd824b26 ignore generated header files with OpenCL kernels 2017-09-19 11:24:38 -04:00
f971cd6e45 Merge pull request #647 from rbberger/gpu_opencl_2
Use OpenCL 2.x API in GPU package when supported
2017-09-12 10:56:13 -06:00
84b84c10db Use OpenCL 2.x API in GPU package when supported 2017-09-09 19:53:56 -04:00
dbc60d949f Fixed OpenCL compilation issues with __global keywords 2017-09-06 15:08:02 +02:00
e88ff8d6f9 correct embedded help for GPU lib Install.py 2017-08-23 15:20:30 -04:00
00aafef1a8 Merge pull request #597 from ndtrung81/three-body-short-nlist
Implementing short neighbor lists for three-body gpu styles
2017-08-17 11:31:59 -06:00
6d24be8bb7 whitespace cleanup 2017-08-10 21:14:40 -04:00
30431d4edb rework Install.py for gpu library. make it consistent with other lib folders and support python3 2017-08-08 16:57:27 -04:00
aa60ef6ed8 Cleaned up 3-body kernels, reverted some mistaken changes to vashishta 2017-07-23 00:08:55 -05:00
a71f5a0c20 Enabled again neigh no with tpa > 1 for 3-body gpu styles for backward compatibility, could be slower than neigh no tpa 1 in many cases 2017-07-22 22:57:37 -05:00
3d1d0c58c7 Cleaned up 3-body gpu styles, and fixed a bug for tersoff/zbl/gpu.
There is a unresolved bug for neigh no with tpa > 1 with BaseThree, enforce tpa = 1 for neigh no in BaseThree for now.
2017-07-21 12:08:04 -05:00
296e572e69 better Install.py script for KIM from Ryan Elliott 2017-07-14 16:25:16 -06:00
cdac5f496c Built 3-body short neighbor list for the 3-body kernels using per-pair cutoffs for vashishta gpu style 2017-07-11 00:13:56 -05:00
8c9db3ea00 Built 2-body short neighbor list and used for 2-body kernels in tersoff gpu styles 2017-07-10 23:50:21 -05:00
ea2b01e83b Refactored 3-body gpu styles to remove code duplication 2017-07-08 20:17:31 -05:00
34fe2273f6 Added short neighbor list implementation for tersoff/zbl/gpu and tersoff/mod/gpu 2017-07-08 14:59:48 -05:00
77c60189b8 Minor cleanups for tersoff/gpu 2017-07-08 14:43:53 -05:00
1c6533e53d Working on short neighbor list for tersoff/gpu 2017-07-08 14:15:26 -05:00
68206079da Supported short neighbor lists for 3-body kernels in sw/gpu and vashishta/gpu 2017-07-07 16:47:24 -05:00
d7646aeeed Fixed opencl error 2017-06-14 12:03:47 +02:00
5f9341813d Removed debug output 2017-06-14 10:57:54 +02:00
8441307185 Removed non-general CUDA-dir in makefile 2017-06-14 10:28:46 +02:00
720af5c360 Added vashishta to OpenCL makefile 2017-06-14 10:27:52 +02:00
eeff0b8633 Added vashishta GPU package for NVidia 2017-06-14 10:24:16 +02:00
addd87c0f7 new Section package and start doc pages and build scripts 2017-05-04 11:22:20 -06:00
b8cb80b219 rename files in GPU library from cg_cmm to lj_sdk 2017-04-26 19:46:10 -04:00
cd435c0c58 change references from cg_cmm to lj_sdk and from cmm to sdk 2017-04-26 19:44:25 -04:00
c8cfd53c1b adjust makefiles in lib tree, so they don't return an error exit code on "make clean" targets.
(cherry picked from commit 4117218ca3)
2016-09-06 21:28:55 -04:00
abc5a32c8a gpu lib sync 2016-08-23 15:30:01 -06:00
34f829c5ac git-svn-id: svn://svn.icms.temple.edu/lammps-ro/trunk@15261 f3b2605a-c512-4ea7-a41b-209d697bcdaa 2016-07-05 14:41:13 +00:00
a99c922a14 git-svn-id: svn://svn.icms.temple.edu/lammps-ro/trunk@15249 f3b2605a-c512-4ea7-a41b-209d697bcdaa 2016-07-01 23:29:38 +00:00
9656958169 git-svn-id: svn://svn.icms.temple.edu/lammps-ro/trunk@15248 f3b2605a-c512-4ea7-a41b-209d697bcdaa 2016-07-01 23:27:26 +00:00