ESPResSo
Extensible Simulation Package for Research on Soft Matter Systems
Loading...
Searching...
No Matches
StreamSweepSinglePrecisionCUDA.h
Go to the documentation of this file.
1//======================================================================================================================
2//
3// This file is part of waLBerla. waLBerla is free software: you can
4// redistribute it and/or modify it under the terms of the GNU General Public
5// License as published by the Free Software Foundation, either version 3 of
6// the License, or (at your option) any later version.
7//
8// waLBerla is distributed in the hope that it will be useful, but WITHOUT
9// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
10// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
11// for more details.
12//
13// You should have received a copy of the GNU General Public License along
14// with waLBerla (see COPYING.txt). If not, see <http://www.gnu.org/licenses/>.
15//
16//! \\file StreamSweepSinglePrecisionCUDA.h
17//! \\author pystencils
18//======================================================================================================================
19
20// kernel generated with pystencils v1.3.3, lbmpy v1.3.3,
21// lbmpy_walberla/pystencils_walberla from waLBerla commit
22// b0842e1a493ce19ef1bbb8d2cf382fc343970a7f
23
24#pragma once
25#include "core/DataTypes.h"
26#include "core/logging/Logging.h"
27
28#include "gpu/GPUField.h"
29#include "gpu/GPUWrapper.h"
30
31#include "domain_decomposition/BlockDataID.h"
32#include "domain_decomposition/IBlock.h"
33#include "domain_decomposition/StructuredBlockStorage.h"
34#include "field/SwapableCompare.h"
35#include <set>
36
37#ifdef __GNUC__
38#define RESTRICT __restrict__
39#elif _MSC_VER
40#define RESTRICT __restrict
41#else
42#define RESTRICT
43#endif
44
45#if (defined WALBERLA_CXX_COMPILER_IS_GNU) || \
46 (defined WALBERLA_CXX_COMPILER_IS_CLANG)
47#pragma GCC diagnostic push
48#pragma GCC diagnostic ignored "-Wunused-parameter"
49#pragma GCC diagnostic ignored "-Wreorder"
50#endif
51
52namespace walberla {
53namespace pystencils {
54
56public:
57 StreamSweepSinglePrecisionCUDA(BlockDataID forceID_, BlockDataID pdfsID_,
58 BlockDataID velocityID_)
59 : forceID(forceID_), pdfsID(pdfsID_), velocityID(velocityID_){};
60
62 for (auto p : cache_pdfs_) {
63 delete p;
64 }
65 }
66
67 void run(IBlock *block, gpuStream_t stream = nullptr);
68
69 void runOnCellInterval(const shared_ptr<StructuredBlockStorage> &blocks,
70 const CellInterval &globalCellInterval,
71 cell_idx_t ghostLayers, IBlock *block,
72 gpuStream_t stream = nullptr);
73
74 void operator()(IBlock *block, gpuStream_t stream = nullptr) {
76 }
77
78 static std::function<void(IBlock *)>
79 getSweep(const shared_ptr<StreamSweepSinglePrecisionCUDA> &kernel) {
80 return [kernel](IBlock *b) { kernel->run(b); };
81 }
82
83 static std::function<void(IBlock *, gpuStream_t)> getSweepOnCellInterval(
84 const shared_ptr<StreamSweepSinglePrecisionCUDA> &kernel,
85 const shared_ptr<StructuredBlockStorage> &blocks,
86 const CellInterval &globalCellInterval, cell_idx_t ghostLayers = 1) {
87 return [kernel, blocks, globalCellInterval,
88 ghostLayers](IBlock *b, gpuStream_t stream = nullptr) {
89 kernel->runOnCellInterval(blocks, globalCellInterval, ghostLayers, b,
90 stream);
91 };
92 }
93
94 std::function<void(IBlock *)> getSweep(gpuStream_t stream = nullptr) {
95 return [this, stream](IBlock *b) { this->run(b, stream); };
96 }
97
98 std::function<void(IBlock *)>
99 getSweepOnCellInterval(const shared_ptr<StructuredBlockStorage> &blocks,
100 const CellInterval &globalCellInterval,
101 cell_idx_t ghostLayers = 1,
102 gpuStream_t stream = nullptr) {
103 return [this, blocks, globalCellInterval, ghostLayers, stream](IBlock *b) {
104 this->runOnCellInterval(blocks, globalCellInterval, ghostLayers, b,
105 stream);
106 };
107 }
108
109 void configure(const shared_ptr<StructuredBlockStorage> &blocks,
110 IBlock *block) {}
111
112 BlockDataID forceID;
113 BlockDataID pdfsID;
114 BlockDataID velocityID;
115
116private:
117 std::set<gpu::GPUField<float> *,
118 field::SwapableCompare<gpu::GPUField<float> *>>
119 cache_pdfs_;
120};
121
122} // namespace pystencils
123} // namespace walberla
124
125#if (defined WALBERLA_CXX_COMPILER_IS_GNU) || \
126 (defined WALBERLA_CXX_COMPILER_IS_CLANG)
127#pragma GCC diagnostic pop
128#endif
static std::function< void(IBlock *, gpuStream_t)> getSweepOnCellInterval(const shared_ptr< StreamSweepSinglePrecisionCUDA > &kernel, const shared_ptr< StructuredBlockStorage > &blocks, const CellInterval &globalCellInterval, cell_idx_t ghostLayers=1)
std::function< void(IBlock *)> getSweep(gpuStream_t stream=nullptr)
std::function< void(IBlock *)> getSweepOnCellInterval(const shared_ptr< StructuredBlockStorage > &blocks, const CellInterval &globalCellInterval, cell_idx_t ghostLayers=1, gpuStream_t stream=nullptr)
void configure(const shared_ptr< StructuredBlockStorage > &blocks, IBlock *block)
void runOnCellInterval(const shared_ptr< StructuredBlockStorage > &blocks, const CellInterval &globalCellInterval, cell_idx_t ghostLayers, IBlock *block, gpuStream_t stream=nullptr)
void operator()(IBlock *block, gpuStream_t stream=nullptr)
static std::function< void(IBlock *)> getSweep(const shared_ptr< StreamSweepSinglePrecisionCUDA > &kernel)
StreamSweepSinglePrecisionCUDA(BlockDataID forceID_, BlockDataID pdfsID_, BlockDataID velocityID_)
cudaStream_t stream[1]
CUDA streams for parallel computing on CPU and GPU.
static double * block(double *p, std::size_t index, std::size_t size)
Definition elc.cpp:172
\file PackInfoPdfDoublePrecision.cpp \author pystencils