1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
|
/*************************************************************************
* Copyright (C) 2009 by Emanuele Catalano <catalano@grenoble-inp.fr> *
* Copyright (C) 2009 by Bruno Chareyre <bruno.chareyre@grenoble-inp.fr> *
* *
* This program is free software; it is licensed under the terms of the *
* GNU General Public License v2 or later. See file LICENSE for details. *
*************************************************************************/
#ifdef YADE_CGAL
#ifdef FLOW_ENGINE
#include "FlowEngine_FlowEngineT.hpp"
#include <lib/serialization/EnumSupport.hpp>
// To register properly, we need to first instantiate an intermediate class, then inherit from it with correct class names in YADE_CLASS macro
// The intermediate one would be seen with the name "TemplateFlowEngine" by python, thus it would not work when more than one class are derived, they would all
// be named "TemplateFlowEngine" ...
namespace yade { // Cannot have #include directive inside.
// YADE_ENUM(yade, UseFlowSolver, (GaussSeidel)(CholeskyFactorization)(MulticoreCPUorGPU));
typedef TemplateFlowEngine_FlowEngineT<FlowCellInfo_FlowEngineT, FlowVertexInfo_FlowEngineT> FlowEngineT;
REGISTER_SERIALIZABLE(FlowEngineT);
class FlowEngine : public FlowEngineT {
public:
// clang-format off
YADE_CLASS_BASE_DOC_ATTRS_INIT_CTOR_PY(FlowEngine,FlowEngineT,"An engine to solve flow problem in saturated granular media. Model description can be found in [Chareyre2012a]_ and [Catalano2014a]_. See the example script FluidCouplingPFV/oedometer.py. More documentation to come.\n\n.. note::Multi-threading seems to work fine for Cholesky decomposition, but it fails for the solve phase in which -j1 is the fastest, here we specify thread numbers independently using :yref:`FlowEngine::numFactorizeThreads` and :yref:`FlowEngine::numSolveThreads`. These multhreading settings are only impacting the behaviour of openblas library and are relatively independant of :yref:`FlowEngine::multithread`. However, the settings have to be globally consistent. For instance, :yref:`multithread<FlowEngine::multithread>` =True with yref:`numFactorizeThreads<FlowEngine::numFactorizeThreads>` = yref:`numSolveThreads<FlowEngine::numSolveThreads>` = 4 implies that openblas will mobilize 8 processors at some point. If the system does not have so many procs. it will hurt performance.",
,,
,
//nothing special to define here, we simply re-use FlowEngine methods
//.def("meanVelocity",&PeriodicFlowEngine::meanVelocity,"measure the mean velocity in the period")
)
// clang-format on
DECLARE_LOGGER;
};
REGISTER_SERIALIZABLE(FlowEngine);
YADE_PLUGIN((FlowEngineT));
CREATE_LOGGER(FlowEngine);
YADE_PLUGIN((FlowEngine));
} // namespace yade
#endif //FLOW_ENGINE
#endif /* YADE_CGAL */
|
