ESPResSo
Extensible Simulation Package for Research on Soft Matter Systems
Loading...
Searching...
No Matches
PackInfoVecSinglePrecision.cpp
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 PackInfoVecSinglePrecision.cpp
17//! \\author pystencils
18//======================================================================================================================
19
20// kernel generated with pystencils v1.4+1.ge851f4e, lbmpy v1.4+1.ge9efe34, sympy v1.12.1, lbmpy_walberla/pystencils_walberla from waLBerla commit 007e77e077ad9d22b5eed6f3d3118240993e553c
21
22#include "PackInfoVecSinglePrecision.h"
23#include "core/DataTypes.h"
24#include "core/cell/CellInterval.h"
25#include "stencil/Directions.h"
26
27#if (defined WALBERLA_CXX_COMPILER_IS_GNU) || (defined WALBERLA_CXX_COMPILER_IS_CLANG)
28#pragma GCC diagnostic push
29#pragma GCC diagnostic ignored "-Wfloat-equal"
30#pragma GCC diagnostic ignored "-Wshadow"
31#pragma GCC diagnostic ignored "-Wconversion"
32#pragma GCC diagnostic ignored "-Wunused-variable"
33#endif
34
35namespace walberla {
36namespace pystencils {
37
38using walberla::cell::CellInterval;
39using walberla::stencil::Direction;
40
41namespace internal_05a1eb9a7382e5e7047cdb22e28b6556 {
42static FUNC_PREFIX void pack_SW_BW_W_TW_NW_BS_S_TS_B_C_T_BN_N_TN_SE_BE_E_TE_NE(float *RESTRICT _data_buffer, float *RESTRICT const _data_field, int64_t const _size_field_0, int64_t const _size_field_1, int64_t const _size_field_2, int64_t const _stride_field_0, int64_t const _stride_field_1, int64_t const _stride_field_2, int64_t const _stride_field_3) {
43 for (int64_t ctr_2 = 0; ctr_2 < _size_field_2; ctr_2 += 1) {
44 for (int64_t ctr_1 = 0; ctr_1 < _size_field_1; ctr_1 += 1) {
45 for (int64_t ctr_0 = 0; ctr_0 < _size_field_0; ctr_0 += 1) {
46 _data_buffer[3 * _size_field_0 * _size_field_1 * ctr_2 + 3 * _size_field_0 * ctr_1 + 3 * ctr_0] = _data_field[_stride_field_0 * ctr_0 + _stride_field_1 * ctr_1 + _stride_field_2 * ctr_2];
47 _data_buffer[3 * _size_field_0 * _size_field_1 * ctr_2 + 3 * _size_field_0 * ctr_1 + 3 * ctr_0 + 1] = _data_field[_stride_field_0 * ctr_0 + _stride_field_1 * ctr_1 + _stride_field_2 * ctr_2 + _stride_field_3];
48 _data_buffer[3 * _size_field_0 * _size_field_1 * ctr_2 + 3 * _size_field_0 * ctr_1 + 3 * ctr_0 + 2] = _data_field[_stride_field_0 * ctr_0 + _stride_field_1 * ctr_1 + _stride_field_2 * ctr_2 + 2 * _stride_field_3];
49 }
50 }
51 }
52}
53} // namespace internal_05a1eb9a7382e5e7047cdb22e28b6556
54
55namespace internal_1ccccad4ca561e07a0934cadb07d0fc1 {
56static FUNC_PREFIX void unpack_SW_BW_W_TW_NW_BS_S_TS_B_C_T_BN_N_TN_SE_BE_E_TE_NE(float *RESTRICT const _data_buffer, float *RESTRICT _data_field, int64_t const _size_field_0, int64_t const _size_field_1, int64_t const _size_field_2, int64_t const _stride_field_0, int64_t const _stride_field_1, int64_t const _stride_field_2, int64_t const _stride_field_3) {
57 for (int64_t ctr_2 = 0; ctr_2 < _size_field_2; ctr_2 += 1) {
58 for (int64_t ctr_1 = 0; ctr_1 < _size_field_1; ctr_1 += 1) {
59 for (int64_t ctr_0 = 0; ctr_0 < _size_field_0; ctr_0 += 1) {
60 _data_field[_stride_field_0 * ctr_0 + _stride_field_1 * ctr_1 + _stride_field_2 * ctr_2] = _data_buffer[3 * _size_field_0 * _size_field_1 * ctr_2 + 3 * _size_field_0 * ctr_1 + 3 * ctr_0];
61 _data_field[_stride_field_0 * ctr_0 + _stride_field_1 * ctr_1 + _stride_field_2 * ctr_2 + _stride_field_3] = _data_buffer[3 * _size_field_0 * _size_field_1 * ctr_2 + 3 * _size_field_0 * ctr_1 + 3 * ctr_0 + 1];
62 _data_field[_stride_field_0 * ctr_0 + _stride_field_1 * ctr_1 + _stride_field_2 * ctr_2 + 2 * _stride_field_3] = _data_buffer[3 * _size_field_0 * _size_field_1 * ctr_2 + 3 * _size_field_0 * ctr_1 + 3 * ctr_0 + 2];
63 }
64 }
65 }
66}
67} // namespace internal_1ccccad4ca561e07a0934cadb07d0fc1
68
69void PackInfoVecSinglePrecision::pack(Direction dir, unsigned char *byte_buffer, IBlock *block) const {
70 float *buffer = reinterpret_cast<float *>(byte_buffer);
71
72 auto field = block->getData<field::GhostLayerField<float, 3>>(fieldID);
73
74 CellInterval ci;
75 field->getSliceBeforeGhostLayer(dir, ci, 1, false);
76
77 switch (dir) {
78 case stencil::SW:
79 case stencil::BW:
80 case stencil::W:
81 case stencil::TW:
82 case stencil::NW:
83 case stencil::BS:
84 case stencil::S:
85 case stencil::TS:
86 case stencil::B:
87 case stencil::C:
88 case stencil::T:
89 case stencil::BN:
90 case stencil::N:
91 case stencil::TN:
92 case stencil::SE:
93 case stencil::BE:
94 case stencil::E:
95 case stencil::TE:
96 case stencil::NE: {
97 float *RESTRICT _data_buffer = buffer;
98 WALBERLA_ASSERT_GREATER_EQUAL(ci.xMin(), -int_c(field->nrOfGhostLayers()))
99 WALBERLA_ASSERT_GREATER_EQUAL(ci.yMin(), -int_c(field->nrOfGhostLayers()))
100 WALBERLA_ASSERT_GREATER_EQUAL(ci.zMin(), -int_c(field->nrOfGhostLayers()))
101 float *RESTRICT const _data_field = field->dataAt(ci.xMin(), ci.yMin(), ci.zMin(), 0);
102 WALBERLA_ASSERT_GREATER_EQUAL(field->xSizeWithGhostLayer(), int64_t(int64_c(ci.xSize()) + 0))
103 const int64_t _size_field_0 = int64_t(int64_c(ci.xSize()) + 0);
104 WALBERLA_ASSERT_GREATER_EQUAL(field->ySizeWithGhostLayer(), int64_t(int64_c(ci.ySize()) + 0))
105 const int64_t _size_field_1 = int64_t(int64_c(ci.ySize()) + 0);
106 WALBERLA_ASSERT_GREATER_EQUAL(field->zSizeWithGhostLayer(), int64_t(int64_c(ci.zSize()) + 0))
107 const int64_t _size_field_2 = int64_t(int64_c(ci.zSize()) + 0);
108 const int64_t _stride_field_0 = int64_t(field->xStride());
109 const int64_t _stride_field_1 = int64_t(field->yStride());
110 const int64_t _stride_field_2 = int64_t(field->zStride());
111 const int64_t _stride_field_3 = int64_t(1 * int64_t(field->fStride()));
112 internal_05a1eb9a7382e5e7047cdb22e28b6556::pack_SW_BW_W_TW_NW_BS_S_TS_B_C_T_BN_N_TN_SE_BE_E_TE_NE(_data_buffer, _data_field, _size_field_0, _size_field_1, _size_field_2, _stride_field_0, _stride_field_1, _stride_field_2, _stride_field_3);
113 break;
114 }
115
116 default:
117 WALBERLA_ASSERT(false);
118 }
119}
120
121void PackInfoVecSinglePrecision::unpack(Direction dir, unsigned char *byte_buffer, IBlock *block) const {
122 float *buffer = reinterpret_cast<float *>(byte_buffer);
123
124 auto field = block->getData<field::GhostLayerField<float, 3>>(fieldID);
125
126 CellInterval ci;
127 field->getGhostRegion(dir, ci, 1, false);
128 auto communciationDirection = stencil::inverseDir[dir];
129
130 switch (communciationDirection) {
131 case stencil::SW:
132 case stencil::BW:
133 case stencil::W:
134 case stencil::TW:
135 case stencil::NW:
136 case stencil::BS:
137 case stencil::S:
138 case stencil::TS:
139 case stencil::B:
140 case stencil::C:
141 case stencil::T:
142 case stencil::BN:
143 case stencil::N:
144 case stencil::TN:
145 case stencil::SE:
146 case stencil::BE:
147 case stencil::E:
148 case stencil::TE:
149 case stencil::NE: {
150 float *RESTRICT const _data_buffer = buffer;
151 WALBERLA_ASSERT_GREATER_EQUAL(ci.xMin(), -int_c(field->nrOfGhostLayers()))
152 WALBERLA_ASSERT_GREATER_EQUAL(ci.yMin(), -int_c(field->nrOfGhostLayers()))
153 WALBERLA_ASSERT_GREATER_EQUAL(ci.zMin(), -int_c(field->nrOfGhostLayers()))
154 float *RESTRICT _data_field = field->dataAt(ci.xMin(), ci.yMin(), ci.zMin(), 0);
155 WALBERLA_ASSERT_GREATER_EQUAL(field->xSizeWithGhostLayer(), int64_t(int64_c(ci.xSize()) + 0))
156 const int64_t _size_field_0 = int64_t(int64_c(ci.xSize()) + 0);
157 WALBERLA_ASSERT_GREATER_EQUAL(field->ySizeWithGhostLayer(), int64_t(int64_c(ci.ySize()) + 0))
158 const int64_t _size_field_1 = int64_t(int64_c(ci.ySize()) + 0);
159 WALBERLA_ASSERT_GREATER_EQUAL(field->zSizeWithGhostLayer(), int64_t(int64_c(ci.zSize()) + 0))
160 const int64_t _size_field_2 = int64_t(int64_c(ci.zSize()) + 0);
161 const int64_t _stride_field_0 = int64_t(field->xStride());
162 const int64_t _stride_field_1 = int64_t(field->yStride());
163 const int64_t _stride_field_2 = int64_t(field->zStride());
164 const int64_t _stride_field_3 = int64_t(1 * int64_t(field->fStride()));
165 internal_1ccccad4ca561e07a0934cadb07d0fc1::unpack_SW_BW_W_TW_NW_BS_S_TS_B_C_T_BN_N_TN_SE_BE_E_TE_NE(_data_buffer, _data_field, _size_field_0, _size_field_1, _size_field_2, _stride_field_0, _stride_field_1, _stride_field_2, _stride_field_3);
166 break;
167 }
168
169 default:
170 WALBERLA_ASSERT(false);
171 }
172}
173
174uint_t PackInfoVecSinglePrecision::size(stencil::Direction dir, const IBlock *block) const {
175 auto field = block->getData<field::GhostLayerField<float, 3>>(fieldID);
176
177 CellInterval ci;
178 field->getGhostRegion(dir, ci, 1, false);
179
180 uint_t elementsPerCell = uint_t{0u};
181
182 switch (dir) {
183 case stencil::SW:
184 case stencil::BW:
185 case stencil::W:
186 case stencil::TW:
187 case stencil::NW:
188 case stencil::BS:
189 case stencil::S:
190 case stencil::TS:
191 case stencil::B:
192 case stencil::C:
193 case stencil::T:
194 case stencil::BN:
195 case stencil::N:
196 case stencil::TN:
197 case stencil::SE:
198 case stencil::BE:
199 case stencil::E:
200 case stencil::TE:
201 case stencil::NE:
202 elementsPerCell = uint_t{3u};
203 break;
204
205 default:
206 elementsPerCell = uint_t{0u};
207 }
208 return ci.numCells() * elementsPerCell * sizeof(float);
209}
210
211} // namespace pystencils
212} // namespace walberla
FUNC_PREFIX
#define FUNC_PREFIX
\file AdvectiveFluxKernel_double_precision.cpp \author pystencils
Definition AdvectiveFluxKernel_double_precision.cpp:28
RESTRICT
#define RESTRICT
\file AdvectiveFluxKernel_double_precision.h \author pystencils
Definition AdvectiveFluxKernel_double_precision.h:40
walberla::pystencils::PackInfoVecSinglePrecision::size
uint_t size(stencil::Direction dir, const IBlock *block) const
Definition PackInfoVecSinglePrecision.cpp:174
walberla::pystencils::PackInfoVecSinglePrecision::pack
void pack(stencil::Direction dir, unsigned char *buffer, IBlock *block) const
Definition PackInfoVecSinglePrecision.cpp:69
walberla::pystencils::PackInfoVecSinglePrecision::unpack
void unpack(stencil::Direction dir, unsigned char *buffer, IBlock *block) const
Definition PackInfoVecSinglePrecision.cpp:121
stream
cudaStream_t stream[1]
CUDA streams for parallel computing on CPU and GPU.
Definition common_cuda.cu:34
block
static double * block(double *p, std::size_t index, std::size_t size)
Definition elc.cpp:177
walberla::pystencils::internal_05a1eb9a7382e5e7047cdb22e28b6556::pack_SW_BW_W_TW_NW_BS_S_TS_B_C_T_BN_N_TN_SE_BE_E_TE_NE
static FUNC_PREFIX void pack_SW_BW_W_TW_NW_BS_S_TS_B_C_T_BN_N_TN_SE_BE_E_TE_NE(double *RESTRICT _data_buffer, double *RESTRICT const _data_field, int64_t const _size_field_0, int64_t const _size_field_1, int64_t const _size_field_2, int64_t const _stride_field_0, int64_t const _stride_field_1, int64_t const _stride_field_2, int64_t const _stride_field_3)
Definition PackInfoVecDoublePrecision.cpp:42
walberla::pystencils::internal_1ccccad4ca561e07a0934cadb07d0fc1::unpack_SW_BW_W_TW_NW_BS_S_TS_B_C_T_BN_N_TN_SE_BE_E_TE_NE
static FUNC_PREFIX void unpack_SW_BW_W_TW_NW_BS_S_TS_B_C_T_BN_N_TN_SE_BE_E_TE_NE(double *RESTRICT const _data_buffer, double *RESTRICT _data_field, int64_t const _size_field_0, int64_t const _size_field_1, int64_t const _size_field_2, int64_t const _stride_field_0, int64_t const _stride_field_1, int64_t const _stride_field_2, int64_t const _stride_field_3)
Definition PackInfoVecDoublePrecision.cpp:56
walberla
\file PackInfoPdfDoublePrecision.cpp \author pystencils
Definition EKWalberla.hpp:38