Commit Graph

349 Commits

Author SHA1 Message Date
jpekkila
230230ead9 Added a script for preprocessing the device files. Useful for inspecting whether the DSL code is generated correctly. 2019-08-27 18:15:30 +03:00
jpekkila
20138263f4 The previous attempt (dsl_feature_completeness_2019-08-23) to enable arbitrary kernel functions was a failure: we get significant performance loss (25-100%) if step_number is not passed as a template parameter to the integration kernel. Apparently the CUDA compiler cannot perform some optimizations if there is a if/else construct in a performance-critical part which cannot be evaluated at compile time. This branch keeps step_number as a template parameter but takes rest of the user parameters as uniforms (dt is no longer passed as a function parameter but as an uniform with the DSL instead). 2019-08-27 17:36:33 +03:00
jpekkila
9fa39b66fc Reverted an accidentally modified astaroth.conf to the same version as in master 2019-08-26 18:33:50 +03:00
jpekkila
022e46f2e7 Merge branch 'master' into dsl_parameter_overhaul_2019-08-19 2019-08-23 13:13:57 +03:00
jpekkila
f6040f89dc Added acPrintMeshInfo for printing all mesh parameters 2019-08-21 16:24:48 +03:00
jpekkila
d52e002e5d Made Astaroth Standalone a library component (still works as before but can be included in other projects which need f.ex. autotesting) 2019-08-21 16:18:48 +03:00
jpekkila
5867ff4b3e Stashing MPItest changes 2019-08-21 16:16:12 +03:00
jpekkila
39dcda4a04 Made warnings about unused functions go away (this is intended functionality and not all programs will use all types of device constants, thus unnecessary warning) 2019-08-21 14:28:46 +03:00
jpekkila
5d2b658fb0 Autoformatted the DSL files 2019-08-20 18:41:26 +03:00
jpekkila
73d393e419 Changed order for linking the MPI library to work around cmake error on CMP0004 2019-08-20 18:40:38 +03:00
jpekkila
5b7408eb55 User config param overhaul complete, works. If I haven't missed anything, all fields and user parameters, and everything related to simulation can now be declared with the DSL. The only thing that you need to do is to fill the declared symbols with data, like with OpenGL and GLSL. 2019-08-19 18:43:16 +03:00
jpekkila
51cf1f1068 The C header is now generated from the DSL, stashing the changes just to be sure since I might overwrite something when updating the compilation scripts to work with this new scheme 2019-08-19 18:19:28 +03:00
jpekkila
d801ebdd41 Now parameters and vertexbuffers (fields) can be declared with the DSL only. TODO: translation from the DSL header to C 2019-08-19 17:35:03 +03:00
jpekkila
bcdd827a4f Added a proper declarations for all user-specified uniform. Note: built-in uniforms are not correctly translated into CUDA 2019-08-19 17:05:56 +03:00
jpekkila
0208d55e4e Moved STENCIL_ORDER and NGHOST out of user-defined parameter as these are actually internal defines used to configure the built-in functions. Additionally, renamed all explicitly declared uniforms from dsx -> AC_dsx in the DSL in preparation for having clear connection between DSL uniforms and the library parameter handles created by the user (AcRealParam etc) 2019-08-19 16:40:47 +03:00
jpekkila
c98b74563c Added a comment 2019-08-19 16:18:24 +03:00
jpekkila
b316e51267 Added preliminary code for generating C headers with the DSL 2019-08-19 16:16:26 +03:00
jpekkila
787363226b Added functions for loading int, int3, scalar and vector constants to the device layer (acDeviceLoad...Constant) 2019-08-19 15:28:16 +03:00
jpekkila
41805dcb68 Added some error checking for the case where user supplies an incomplete meshinfo to acDeviceLoadMeshInfo 2019-08-19 15:17:51 +03:00
jpekkila
598799d7c3 Added a new function to the device interface: acDeviceLoadMeshInfo 2019-08-19 15:14:00 +03:00
jpekkila
e89897985e Battled with math.h and cmath. We probably should move from C standard libraries to C++ ones internally (in places which are not visible via the interface) 2019-08-19 14:02:30 +03:00
jpekkila
6d4d53342e Removed old comments 2019-08-15 11:14:52 +03:00
jpekkila
aa45ce04de Made the linear algebra functions used in forcing.cc static to avoid collisions with the functions defined in math_utils.h 2019-08-15 11:09:40 +03:00
jpekkila
36fea70560 Moved basic built-in functions for vector operations to math_utils.h from integration.cuh so that they are shared with the CPU and GPU 2019-08-15 11:04:22 +03:00
jpekkila
d5b2e5bb42 Added placeholders for new built-in variables in the DSL. Also overloads to DCONST_INT etc. Naming still pending and old DCONST_REAL etc calls still work. 2019-08-12 14:05:35 +03:00
jpekkila
b8c4d07de2 Removed unnecessary comments 2019-08-12 13:31:24 +03:00
jpekkila
e027f7e548 Removed grid_n in astaroth.cu and replaced it with the new acNodeQueryDeviceConfiguration call 2019-08-12 13:25:47 +03:00
jpekkila
3369d8efec Added a missing include 2019-08-12 11:44:27 +03:00
jpekkila
bba9ec7c3b Implemented acNodeQueryDeviceConfiguration 2019-08-12 11:40:38 +03:00
jpekkila
b5daf22c26 Added interface function acSynchronizeMesh 2019-08-12 10:25:05 +03:00
jpekkila
8bbb2cd5df Now prints device info before trying to run the dummy kernel 2019-08-12 09:46:37 +03:00
jpekkila
fdadd463b7 Included the user-defined header after the definition of AcReal to make it available if needed. 2019-08-09 17:11:21 +03:00
jpekkila
bc0276ba2d Merged in dsl_syntax_revision_08008 (pull request #5)
Dsl syntax revision 08008

Approved-by: Miikka Väisälä <mvaisala@asiaa.sinica.edu.tw>
2019-08-09 05:16:37 +00:00
jpekkila
a7a7b535d6 The code generator now generates more efficient CUDA: built-in parameters are passed as const references without copying 2019-08-08 21:35:07 +03:00
jpekkila
a172abeb1f Modified the other sps files to use the new syntax. Though does not compile since there are some old/very old changes in the DSL that have not been updated to these files (f.ex. RK macro does not exist anymore, it's currently rk3) 2019-08-08 21:25:45 +03:00
jpekkila
67358fa75f Removed acc/samples as unmaintained 2019-08-08 21:22:34 +03:00
jpekkila
b53cabbc44 Made the DSL syntax less confusing: Input and output arrays are now ScalarField and VectorFields instead of scalars and vectors. C++ initializers are now also possible, removing the need to declare Fields as int or int3 which was very confusing, like "what, you assing an int value to a real, what the &^%@?" 2019-08-08 21:07:36 +03:00
jpekkila
5397495496 Added acLoadWithOffset 2019-08-08 20:43:01 +03:00
jpekkila
e79e1207f2 Added a function for checking whether CUDA-capable devices are available 2019-08-08 20:35:02 +03:00
jpekkila
8a9099d75e Added missing functions to fix backwards compatibility with the version interfaced with Pencil Code 2019-08-08 19:49:57 +03:00
jpekkila
168cdc9109 Fixed a float/long double conversion that some complainers... um compilers complain about 2019-08-08 18:12:24 +03:00
jpekkila
af24f17d1f Merged in node_device_interface_revision_07-23 (pull request #4)
Node device interface revision 07 23
2019-08-08 02:52:31 +00:00
jpekkila
322cdce52c Added some new comments + some helpful old comments from a time before the interface revision 2019-08-07 20:05:54 +03:00
jpekkila
2b3f9d75af Ensured that acBoundcondStep is called everywhere in the program before acStore 2019-08-07 19:20:40 +03:00
jpekkila
d8eb2723b4 Added an acBoundconds() call before acStore in autotest.cc 2019-08-07 19:10:04 +03:00
jpekkila
1525e0603f Added some preliminary pragma omps and verified that acIntegrate works as it should. 2019-08-07 19:08:52 +03:00
jpekkila
c2bd5ae3e6 Simplified the optimized multi-GPU integration function 2019-08-07 18:17:03 +03:00
jpekkila
fd94b6321d Renamed globalGrid.n to globalGridN 2019-08-07 18:16:34 +03:00
jpekkila
10200e4dd5 Merge branch 'master' into node_device_interface_revision_07-23 2019-08-07 16:25:33 +03:00
jpekkila
5432d20b1b Removed a depreceated auto-optimization script. The optimization is done during acInit() at runtime instead 2019-08-07 16:25:03 +03:00