Clement Jamin
2dc6c405b8
Drop the use of the TBB malloc proxy, use TBB allocators manually instead
...
The TBB's "malloc proxy", which allows to replace all calls to the standard
allocators by calls to the TBB scalable allocator, is not available on MacOS.
Instead, we use the TBB allocators manually, where relevant.
2013-06-27 15:29:47 +02:00
Clement Jamin
bc360c8b13
Changes for GCC compilation + removed some trailing spaces
2013-06-07 17:40:53 +02:00
Clement Jamin
440a654a58
Added a concurrency-safe version of DT3::find_conflicts + better doc
2013-06-05 17:23:19 +02:00
Clement Jamin
b43f28c6ab
Renamed p_could_lock_zone + use NULL instead of 0 + CGAL_TRIANGULATION_3_PROFILING
2013-06-05 13:44:34 +02:00
Clement Jamin
dbe1c9c597
Reversed the "facet_not_in_its_cz" variables by removing the "not"
2013-05-28 15:56:22 +02:00
Clement Jamin
eeeb6cc841
Smaller tag names (continued)
2013-05-28 15:42:39 +02:00
Clement Jamin
127926aa14
Proof-reading + minor changes
...
Renamed unlock_all_tls_locked_locations => unlock_all_points_locked_by_this_thread
2013-05-28 15:07:20 +02:00
Clement Jamin
9aefdbe939
Replaced is_base_of by is_convertible.
2013-05-23 17:49:24 +02:00
Clement Jamin
c471fb442d
Untabify + removed useless spaces
2013-05-21 11:13:36 +02:00
Clement Jamin
3a4caaaa1c
Changed the place of the erase counters.
...
Changed Compact_container_strategy_XXX::Uses_erase_counter type + moved the erase counters from Triangulation_ds_xxx_base_3 to Mesh_xxx_base_3.
2013-04-16 15:31:36 +02:00
Clement Jamin
2a7c402c07
Regular_triangulation_3 is now parallel (insert/remove)
2013-04-10 18:08:53 +02:00
Clement Jamin
6a85889a1e
New constructor + minor changes
2013-04-10 18:07:14 +02:00
Clement Jamin
ef756afafd
Missing "dummy function" for the sequential case
2013-04-03 17:35:24 +02:00
Clement Jamin
1cc98e9abd
Updated doc + code clean-up
2013-04-03 17:34:21 +02:00
Clement Jamin
e6b33376bb
Avoid multiple calls to tbb::enumerable_thread_specific::local()
2013-03-29 17:16:26 +01:00
Clement Jamin
4399f8218b
Untabify (missed some of them last time)
2013-03-28 10:13:32 +01:00
Clement Jamin
8c19195bcb
Untabify
2013-03-27 15:32:10 +01:00
Clement Jamin
cc9dde002e
Untabify
2013-03-27 15:30:46 +01:00
Clement Jamin
c3e9699887
Clean-up + better "profiling"
2013-03-27 14:39:10 +01:00
Clement Jamin
de1f43e30b
The remove is now fully parallel.
...
The remove operations are done in parallel as long as the dimension stays = 3, then the few remaining points are removed sequentially.
2013-03-27 13:47:40 +01:00
Clement Jamin
57d84ecb24
New constructor for Tri_3 and TDS + new way of testing "dim down" in the 3D case
...
New constructor taking 4 non-coplanar vertices, for faster construction of very small triangulations => a 3D triangulation is directly constructed.
The new "test_dim_down_using_incident_cells_3" is used in the 3D and will be used by the parallel version.
2013-03-26 19:00:37 +01:00
Clement Jamin
b83b159c2f
Fix so that we don't require xx_erase_counter functions anymore when sequential
2013-03-14 17:45:19 +01:00
Clement Jamin
1f0e994a00
Renamed a few functions + tabs to spaces conversion + benchmark output + minor changes
2013-03-11 15:46:27 +01:00
Clement Jamin
a7cfd06598
Replaced the unused pair<bool,int> returned by try_lock by a simple bool.
2013-03-08 16:47:37 +01:00
Clement Jamin
3eae2927e7
Merge branch 'Mesh_3-parallel-cjamin-old' into Mesh_3-parallel-cjamin
2013-03-04 15:50:57 +01:00
Clement Jamin
a8d38a5fed
Code clean-up
2013-02-28 19:46:06 +01:00
Laurent Rineau
bc7e2a1d71
Merge remote-tracking branch 'origin/Mesh_3-improvements-GF' into Mesh_3-improvements-GF-new
...
Mesh_3-improvements-GF-new is fresh fork of master
Conflicts:
.gitattributes
AABB_tree/include/CGAL/AABB_tree.h
GraphicsView/src/CGAL_Qt4/DemosMainWindow.cpp
Mesh_3/include/CGAL/Mesh_3/C3T3_helpers.h
Mesh_3/include/CGAL/Mesh_3/Refine_facets_3.h
Mesh_3/include/CGAL/Polyhedral_mesh_domain_3.h
2013-02-28 17:10:16 +01:00
Laurent Rineau
bd0930fc26
Fix segfaults in T3, caused by the use of CGAL_assume
...
I wonder why the compiler created buggy code. Anyway, I found easy
workarounds, without reintroducing the warnings.
2013-02-22 16:15:04 +01:00
Laurent Rineau
66c5d23154
Use CGAL_assume to remove warnings
2013-02-20 18:55:43 +01:00
Clement Jamin
a29a847e55
Some tests with parallel removal of vertices.
2013-02-20 17:39:15 +01:00
Clement Jamin
44b8217d54
New/modified constructors for triangulations.
...
Can now take a lock data structure as parameter.
2013-02-19 10:39:39 +01:00
Clement Jamin
864f4eca81
Moved Spatial_grid_lock_data_structure_3.h to STL_Extension
2013-02-18 13:09:12 +01:00
Clement Jamin
8d5352e15f
Better grid lock data structure.
...
There is just one grid lock data structure class now, with several specialization.
Moreover, now the type of the lock data structure is only defined in the triangulation.
2013-02-18 12:58:57 +01:00
Clement Jamin
0c80db26d4
Forgot to remove an #include
2013-02-18 11:16:45 +01:00
Clement Jamin
bce43fbb8e
Improved the way parallelism is handle in Mesh_3/Triangulation_3
...
Added a better support for parallelism in Triangulation_3.
Simplified how to enable/disable concurrency in Mesh_3.
Moved the Compact_container stategies to a new file.
2013-02-15 18:07:43 +01:00
Andreas Fabri
8b5b56adb1
make index an unsigned int
2013-02-15 11:30:08 +01:00
Clement Jamin
eb2ed63d8b
Changed the way we activate concurrency in Triangulation_3 + Code clean-up
2013-02-11 19:12:30 +01:00
Andreas Fabri
afeea1e124
replace size_t with int in loops over the vertices of a cell
2013-02-07 17:26:32 +01:00
Clement Jamin
c4859681e3
Removed the "localization ids" related code.
...
It was an old, unused and slow.
2013-02-06 18:57:36 +01:00
Clement Jamin
cb89562205
Removed the CGAL_MESH_3_ACTIVATE_GRID_INDEX_CACHE_IN_VERTEX macro.
...
It was buggy and useless.
2013-02-06 18:10:35 +01:00
Clement Jamin
5938113aff
Moved some code because there was a problem with some #ifdef
2013-02-06 14:30:04 +01:00
Clement Jamin
dc5e90b8cc
Clean-up + some debugging functions + comments
2013-02-02 06:53:43 +01:00
Clement Jamin
2ee92ca346
Missing #ifdef
2013-01-30 17:30:31 +01:00
Clement Jamin
ad4a9dcaf6
New is_element_locked_by_this_thread for testing purposes
2013-01-29 19:03:07 +01:00
Clement Jamin
7af597cb4f
Parallel Mesh_3 can now be compiled with GCC
2013-01-17 14:49:04 +01:00
Clement Jamin
3dd7981cf4
The erase counter in vertex_base is only required for the parallel version.
2013-01-15 17:05:24 +01:00
Clement Jamin
44e91351ce
Code clean-up + minor changes/optimizations
2013-01-11 16:40:03 +01:00
Clement Jamin
7efd61b6db
Removed some data race conditions in the parallel perturber.
...
Note: the code needs some clean-up
2013-01-08 08:59:17 +01:00
Clement Jamin
dc7096b1e5
This parallel version of the Perturber seems much better. Needs to be tested...
2012-12-13 20:39:37 +01:00
Philipp Möller
3e98c46535
Correct syntax for uniform initialization
2012-12-12 11:24:34 +01:00