ESPResSo
Extensible Simulation Package for Research on Soft Matter Systems
Loading...
Searching...
No Matches
StreamSweepSinglePrecisionCUDA.cu
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.cpp
17//! \\author pystencils
18//======================================================================================================================
19
20// kernel generated with pystencils v1.3.7, lbmpy v1.3.7, sympy v1.12.1, lbmpy_walberla/pystencils_walberla from waLBerla commit f36fa0a68bae59f0b516f6587ea8fa7c24a41141
21
22#include <cmath>
23
24#include "StreamSweepSinglePrecisionCUDA.h"
25#include "core/DataTypes.h"
26#include "core/Macros.h"
27
28#define FUNC_PREFIX __global__
29
30#if (defined WALBERLA_CXX_COMPILER_IS_GNU) || (defined WALBERLA_CXX_COMPILER_IS_CLANG)
31#pragma GCC diagnostic push
32#pragma GCC diagnostic ignored "-Wfloat-equal"
33#pragma GCC diagnostic ignored "-Wshadow"
34#pragma GCC diagnostic ignored "-Wconversion"
35#pragma GCC diagnostic ignored "-Wunused-variable"
36#endif
37
38#if (defined WALBERLA_CXX_COMPILER_IS_INTEL)
39#pragma warning push
40#pragma warning(disable : 1599)
41#endif
42
43using namespace std;
44
45namespace walberla {
46namespace pystencils {
47
48namespace internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda {
49static FUNC_PREFIX __launch_bounds__(256) void streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda(float *RESTRICT const _data_force, float *RESTRICT const _data_pdfs, float *RESTRICT _data_pdfs_tmp, float *RESTRICT _data_velocity, int64_t const _size_force_0, int64_t const _size_force_1, int64_t const _size_force_2, int64_t const _stride_force_0, int64_t const _stride_force_1, int64_t const _stride_force_2, int64_t const _stride_force_3, int64_t const _stride_pdfs_0, int64_t const _stride_pdfs_1, int64_t const _stride_pdfs_2, int64_t const _stride_pdfs_3, int64_t const _stride_pdfs_tmp_0, int64_t const _stride_pdfs_tmp_1, int64_t const _stride_pdfs_tmp_2, int64_t const _stride_pdfs_tmp_3, int64_t const _stride_velocity_0, int64_t const _stride_velocity_1, int64_t const _stride_velocity_2, int64_t const _stride_velocity_3) {
50 if (blockDim.x * blockIdx.x + threadIdx.x + 1 < _size_force_0 - 1 && blockDim.y * blockIdx.y + threadIdx.y + 1 < _size_force_1 - 1 && blockDim.z * blockIdx.z + threadIdx.z + 1 < _size_force_2 - 1) {
51 const int64_t ctr_0 = blockDim.x * blockIdx.x + threadIdx.x + 1;
52 const int64_t ctr_1 = blockDim.y * blockIdx.y + threadIdx.y + 1;
53 const int64_t ctr_2 = blockDim.z * blockIdx.z + threadIdx.z + 1;
54 const float streamed_0 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_2 * ctr_2];
55 const float streamed_1 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_1 * ctr_1 - _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 + _stride_pdfs_3];
56 const float streamed_2 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 + 2 * _stride_pdfs_3];
57 const float streamed_3 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_2 * ctr_2 + 3 * _stride_pdfs_3];
58 const float streamed_4 = _data_pdfs[_stride_pdfs_0 * ctr_0 - _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_2 * ctr_2 + 4 * _stride_pdfs_3];
59 const float streamed_5 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_2 * ctr_2 - _stride_pdfs_2 + 5 * _stride_pdfs_3];
60 const float streamed_6 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_2 * ctr_2 + _stride_pdfs_2 + 6 * _stride_pdfs_3];
61 const float streamed_7 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 - _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 + 7 * _stride_pdfs_3];
62 const float streamed_8 = _data_pdfs[_stride_pdfs_0 * ctr_0 - _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 - _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 + 8 * _stride_pdfs_3];
63 const float streamed_9 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 + 9 * _stride_pdfs_3];
64 const float streamed_10 = _data_pdfs[_stride_pdfs_0 * ctr_0 - _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 + 10 * _stride_pdfs_3];
65 const float streamed_11 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_1 * ctr_1 - _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 - _stride_pdfs_2 + 11 * _stride_pdfs_3];
66 const float streamed_12 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 - _stride_pdfs_2 + 12 * _stride_pdfs_3];
67 const float streamed_13 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_2 * ctr_2 - _stride_pdfs_2 + 13 * _stride_pdfs_3];
68 const float streamed_14 = _data_pdfs[_stride_pdfs_0 * ctr_0 - _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_2 * ctr_2 - _stride_pdfs_2 + 14 * _stride_pdfs_3];
69 const float streamed_15 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_1 * ctr_1 - _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 + _stride_pdfs_2 + 15 * _stride_pdfs_3];
70 const float streamed_16 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_1 + _stride_pdfs_2 * ctr_2 + _stride_pdfs_2 + 16 * _stride_pdfs_3];
71 const float streamed_17 = _data_pdfs[_stride_pdfs_0 * ctr_0 + _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_2 * ctr_2 + _stride_pdfs_2 + 17 * _stride_pdfs_3];
72 const float streamed_18 = _data_pdfs[_stride_pdfs_0 * ctr_0 - _stride_pdfs_0 + _stride_pdfs_1 * ctr_1 + _stride_pdfs_2 * ctr_2 + _stride_pdfs_2 + 18 * _stride_pdfs_3];
73 const float vel0Term = streamed_10 + streamed_14 + streamed_18 + streamed_4 + streamed_8;
74 const float momdensity_0 = -streamed_13 - streamed_17 - streamed_3 - streamed_7 - streamed_9 + vel0Term;
75 const float vel1Term = streamed_1 + streamed_11 + streamed_15 + streamed_7;
76 const float momdensity_1 = -streamed_10 - streamed_12 - streamed_16 - streamed_2 + streamed_8 - streamed_9 + vel1Term;
77 const float vel2Term = streamed_12 + streamed_13 + streamed_5;
78 const float rho = streamed_0 + streamed_16 + streamed_17 + streamed_2 + streamed_3 + streamed_6 + streamed_9 + vel0Term + vel1Term + vel2Term;
79 const float momdensity_2 = streamed_11 + streamed_14 - streamed_15 - streamed_16 - streamed_17 - streamed_18 - streamed_6 + vel2Term;
80 const float u_0 = momdensity_0 * ((1.0f) / (rho)) + 0.5f * ((1.0f) / (rho)) * _data_force[_stride_force_0 * ctr_0 + _stride_force_1 * ctr_1 + _stride_force_2 * ctr_2];
81 const float u_1 = momdensity_1 * ((1.0f) / (rho)) + 0.5f * ((1.0f) / (rho)) * _data_force[_stride_force_0 * ctr_0 + _stride_force_1 * ctr_1 + _stride_force_2 * ctr_2 + _stride_force_3];
82 const float u_2 = momdensity_2 * ((1.0f) / (rho)) + 0.5f * ((1.0f) / (rho)) * _data_force[_stride_force_0 * ctr_0 + _stride_force_1 * ctr_1 + _stride_force_2 * ctr_2 + 2 * _stride_force_3];
83 _data_velocity[_stride_velocity_0 * ctr_0 + _stride_velocity_1 * ctr_1 + _stride_velocity_2 * ctr_2] = u_0;
84 _data_velocity[_stride_velocity_0 * ctr_0 + _stride_velocity_1 * ctr_1 + _stride_velocity_2 * ctr_2 + _stride_velocity_3] = u_1;
85 _data_velocity[_stride_velocity_0 * ctr_0 + _stride_velocity_1 * ctr_1 + _stride_velocity_2 * ctr_2 + 2 * _stride_velocity_3] = u_2;
86 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2] = streamed_0;
87 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + _stride_pdfs_tmp_3] = streamed_1;
88 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 2 * _stride_pdfs_tmp_3] = streamed_2;
89 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 3 * _stride_pdfs_tmp_3] = streamed_3;
90 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 4 * _stride_pdfs_tmp_3] = streamed_4;
91 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 5 * _stride_pdfs_tmp_3] = streamed_5;
92 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 6 * _stride_pdfs_tmp_3] = streamed_6;
93 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 7 * _stride_pdfs_tmp_3] = streamed_7;
94 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 8 * _stride_pdfs_tmp_3] = streamed_8;
95 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 9 * _stride_pdfs_tmp_3] = streamed_9;
96 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 10 * _stride_pdfs_tmp_3] = streamed_10;
97 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 11 * _stride_pdfs_tmp_3] = streamed_11;
98 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 12 * _stride_pdfs_tmp_3] = streamed_12;
99 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 13 * _stride_pdfs_tmp_3] = streamed_13;
100 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 14 * _stride_pdfs_tmp_3] = streamed_14;
101 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 15 * _stride_pdfs_tmp_3] = streamed_15;
102 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 16 * _stride_pdfs_tmp_3] = streamed_16;
103 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 17 * _stride_pdfs_tmp_3] = streamed_17;
104 _data_pdfs_tmp[_stride_pdfs_tmp_0 * ctr_0 + _stride_pdfs_tmp_1 * ctr_1 + _stride_pdfs_tmp_2 * ctr_2 + 18 * _stride_pdfs_tmp_3] = streamed_18;
105 }
106}
107} // namespace internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda
108
109void StreamSweepSinglePrecisionCUDA::run(IBlock *block, gpuStream_t stream) {
110
111 auto velocity = block->getData<gpu::GPUField<float>>(velocityID);
112 auto pdfs = block->getData<gpu::GPUField<float>>(pdfsID);
113 auto force = block->getData<gpu::GPUField<float>>(forceID);
114 gpu::GPUField<float> *pdfs_tmp;
115 {
116 if (cache_pdfs_.find(block) == cache_pdfs_.end()) {
117 pdfs_tmp = pdfs->cloneUninitialized();
118 cache_pdfs_[block] = pdfs_tmp;
119 } else {
120 pdfs_tmp = cache_pdfs_[block];
121 }
122 }
123
124 WALBERLA_ASSERT_GREATER_EQUAL(-1, -int_c(force->nrOfGhostLayers()))
125 float *RESTRICT const _data_force = force->dataAt(-1, -1, -1, 0);
126 WALBERLA_ASSERT_EQUAL(force->layout(), field::fzyx)
127 WALBERLA_ASSERT_GREATER_EQUAL(-1, -int_c(pdfs->nrOfGhostLayers()))
128 float *RESTRICT const _data_pdfs = pdfs->dataAt(-1, -1, -1, 0);
129 WALBERLA_ASSERT_EQUAL(pdfs->layout(), field::fzyx)
130 WALBERLA_ASSERT_GREATER_EQUAL(-1, -int_c(pdfs_tmp->nrOfGhostLayers()))
131 float *RESTRICT _data_pdfs_tmp = pdfs_tmp->dataAt(-1, -1, -1, 0);
132 WALBERLA_ASSERT_EQUAL(pdfs_tmp->layout(), field::fzyx)
133 WALBERLA_ASSERT_GREATER_EQUAL(-1, -int_c(velocity->nrOfGhostLayers()))
134 float *RESTRICT _data_velocity = velocity->dataAt(-1, -1, -1, 0);
135 WALBERLA_ASSERT_EQUAL(velocity->layout(), field::fzyx)
136 WALBERLA_ASSERT_GREATER_EQUAL(force->xSizeWithGhostLayer(), int64_t(int64_c(force->xSize()) + 2))
137 const int64_t _size_force_0 = int64_t(int64_c(force->xSize()) + 2);
138 WALBERLA_ASSERT_EQUAL(force->layout(), field::fzyx)
139 WALBERLA_ASSERT_GREATER_EQUAL(force->ySizeWithGhostLayer(), int64_t(int64_c(force->ySize()) + 2))
140 const int64_t _size_force_1 = int64_t(int64_c(force->ySize()) + 2);
141 WALBERLA_ASSERT_EQUAL(force->layout(), field::fzyx)
142 WALBERLA_ASSERT_GREATER_EQUAL(force->zSizeWithGhostLayer(), int64_t(int64_c(force->zSize()) + 2))
143 const int64_t _size_force_2 = int64_t(int64_c(force->zSize()) + 2);
144 WALBERLA_ASSERT_EQUAL(force->layout(), field::fzyx)
145 const int64_t _stride_force_0 = int64_t(force->xStride());
146 const int64_t _stride_force_1 = int64_t(force->yStride());
147 const int64_t _stride_force_2 = int64_t(force->zStride());
148 const int64_t _stride_force_3 = int64_t(1 * int64_t(force->fStride()));
149 const int64_t _stride_pdfs_0 = int64_t(pdfs->xStride());
150 const int64_t _stride_pdfs_1 = int64_t(pdfs->yStride());
151 const int64_t _stride_pdfs_2 = int64_t(pdfs->zStride());
152 const int64_t _stride_pdfs_3 = int64_t(1 * int64_t(pdfs->fStride()));
153 const int64_t _stride_pdfs_tmp_0 = int64_t(pdfs_tmp->xStride());
154 const int64_t _stride_pdfs_tmp_1 = int64_t(pdfs_tmp->yStride());
155 const int64_t _stride_pdfs_tmp_2 = int64_t(pdfs_tmp->zStride());
156 const int64_t _stride_pdfs_tmp_3 = int64_t(1 * int64_t(pdfs_tmp->fStride()));
157 const int64_t _stride_velocity_0 = int64_t(velocity->xStride());
158 const int64_t _stride_velocity_1 = int64_t(velocity->yStride());
159 const int64_t _stride_velocity_2 = int64_t(velocity->zStride());
160 const int64_t _stride_velocity_3 = int64_t(1 * int64_t(velocity->fStride()));
161 dim3 _block(uint32_c(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)), uint32_c(((1024 < ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))) ? 1024 : ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))), uint32_c(((64 < ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))) ? 64 : ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))))));
162 dim3 _grid(uint32_c(((_size_force_0 - 2) % (((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)) == 0 ? (int64_t)(_size_force_0 - 2) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)) : ((int64_t)(_size_force_0 - 2) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))) + 1)), uint32_c(((_size_force_1 - 2) % (((1024 < ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))) ? 1024 : ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))) == 0 ? (int64_t)(_size_force_1 - 2) / (int64_t)(((1024 < ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))) ? 1024 : ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))) : ((int64_t)(_size_force_1 - 2) / (int64_t)(((1024 < ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))) ? 1024 : ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) + 1)), uint32_c(((_size_force_2 - 2) % (((64 < ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))) ? 64 : ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))))) == 0 ? (int64_t)(_size_force_2 - 2) / (int64_t)(((64 < ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))) ? 64 : ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))))) : ((int64_t)(_size_force_2 - 2) / (int64_t)(((64 < ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))) ? 64 : ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))))) + 1)));
163 internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda<<<_grid, _block, 0, stream>>>(_data_force, _data_pdfs, _data_pdfs_tmp, _data_velocity, _size_force_0, _size_force_1, _size_force_2, _stride_force_0, _stride_force_1, _stride_force_2, _stride_force_3, _stride_pdfs_0, _stride_pdfs_1, _stride_pdfs_2, _stride_pdfs_3, _stride_pdfs_tmp_0, _stride_pdfs_tmp_1, _stride_pdfs_tmp_2, _stride_pdfs_tmp_3, _stride_velocity_0, _stride_velocity_1, _stride_velocity_2, _stride_velocity_3);
164 pdfs->swapDataPointers(pdfs_tmp);
165}
166
167void StreamSweepSinglePrecisionCUDA::runOnCellInterval(const shared_ptr<StructuredBlockStorage> &blocks, const CellInterval &globalCellInterval, cell_idx_t ghostLayers, IBlock *block, gpuStream_t stream) {
168
169 CellInterval ci = globalCellInterval;
170 CellInterval blockBB = blocks->getBlockCellBB(*block);
171 blockBB.expand(ghostLayers);
172 ci.intersect(blockBB);
173 blocks->transformGlobalToBlockLocalCellInterval(ci, *block);
174 if (ci.empty())
175 return;
176
177 auto velocity = block->getData<gpu::GPUField<float>>(velocityID);
178 auto pdfs = block->getData<gpu::GPUField<float>>(pdfsID);
179 auto force = block->getData<gpu::GPUField<float>>(forceID);
180 gpu::GPUField<float> *pdfs_tmp;
181 {
182 if (cache_pdfs_.find(block) == cache_pdfs_.end()) {
183 pdfs_tmp = pdfs->cloneUninitialized();
184 cache_pdfs_[block] = pdfs_tmp;
185 } else {
186 pdfs_tmp = cache_pdfs_[block];
187 }
188 }
189
190 WALBERLA_ASSERT_GREATER_EQUAL(ci.xMin() - 1, -int_c(force->nrOfGhostLayers()))
191 WALBERLA_ASSERT_GREATER_EQUAL(ci.yMin() - 1, -int_c(force->nrOfGhostLayers()))
192 WALBERLA_ASSERT_GREATER_EQUAL(ci.zMin() - 1, -int_c(force->nrOfGhostLayers()))
193 float *RESTRICT const _data_force = force->dataAt(ci.xMin() - 1, ci.yMin() - 1, ci.zMin() - 1, 0);
194 WALBERLA_ASSERT_EQUAL(force->layout(), field::fzyx)
195 WALBERLA_ASSERT_GREATER_EQUAL(ci.xMin() - 1, -int_c(pdfs->nrOfGhostLayers()))
196 WALBERLA_ASSERT_GREATER_EQUAL(ci.yMin() - 1, -int_c(pdfs->nrOfGhostLayers()))
197 WALBERLA_ASSERT_GREATER_EQUAL(ci.zMin() - 1, -int_c(pdfs->nrOfGhostLayers()))
198 float *RESTRICT const _data_pdfs = pdfs->dataAt(ci.xMin() - 1, ci.yMin() - 1, ci.zMin() - 1, 0);
199 WALBERLA_ASSERT_EQUAL(pdfs->layout(), field::fzyx)
200 WALBERLA_ASSERT_GREATER_EQUAL(ci.xMin() - 1, -int_c(pdfs_tmp->nrOfGhostLayers()))
201 WALBERLA_ASSERT_GREATER_EQUAL(ci.yMin() - 1, -int_c(pdfs_tmp->nrOfGhostLayers()))
202 WALBERLA_ASSERT_GREATER_EQUAL(ci.zMin() - 1, -int_c(pdfs_tmp->nrOfGhostLayers()))
203 float *RESTRICT _data_pdfs_tmp = pdfs_tmp->dataAt(ci.xMin() - 1, ci.yMin() - 1, ci.zMin() - 1, 0);
204 WALBERLA_ASSERT_EQUAL(pdfs_tmp->layout(), field::fzyx)
205 WALBERLA_ASSERT_GREATER_EQUAL(ci.xMin() - 1, -int_c(velocity->nrOfGhostLayers()))
206 WALBERLA_ASSERT_GREATER_EQUAL(ci.yMin() - 1, -int_c(velocity->nrOfGhostLayers()))
207 WALBERLA_ASSERT_GREATER_EQUAL(ci.zMin() - 1, -int_c(velocity->nrOfGhostLayers()))
208 float *RESTRICT _data_velocity = velocity->dataAt(ci.xMin() - 1, ci.yMin() - 1, ci.zMin() - 1, 0);
209 WALBERLA_ASSERT_EQUAL(velocity->layout(), field::fzyx)
210 WALBERLA_ASSERT_GREATER_EQUAL(force->xSizeWithGhostLayer(), int64_t(int64_c(ci.xSize()) + 2))
211 const int64_t _size_force_0 = int64_t(int64_c(ci.xSize()) + 2);
212 WALBERLA_ASSERT_EQUAL(force->layout(), field::fzyx)
213 WALBERLA_ASSERT_GREATER_EQUAL(force->ySizeWithGhostLayer(), int64_t(int64_c(ci.ySize()) + 2))
214 const int64_t _size_force_1 = int64_t(int64_c(ci.ySize()) + 2);
215 WALBERLA_ASSERT_EQUAL(force->layout(), field::fzyx)
216 WALBERLA_ASSERT_GREATER_EQUAL(force->zSizeWithGhostLayer(), int64_t(int64_c(ci.zSize()) + 2))
217 const int64_t _size_force_2 = int64_t(int64_c(ci.zSize()) + 2);
218 WALBERLA_ASSERT_EQUAL(force->layout(), field::fzyx)
219 const int64_t _stride_force_0 = int64_t(force->xStride());
220 const int64_t _stride_force_1 = int64_t(force->yStride());
221 const int64_t _stride_force_2 = int64_t(force->zStride());
222 const int64_t _stride_force_3 = int64_t(1 * int64_t(force->fStride()));
223 const int64_t _stride_pdfs_0 = int64_t(pdfs->xStride());
224 const int64_t _stride_pdfs_1 = int64_t(pdfs->yStride());
225 const int64_t _stride_pdfs_2 = int64_t(pdfs->zStride());
226 const int64_t _stride_pdfs_3 = int64_t(1 * int64_t(pdfs->fStride()));
227 const int64_t _stride_pdfs_tmp_0 = int64_t(pdfs_tmp->xStride());
228 const int64_t _stride_pdfs_tmp_1 = int64_t(pdfs_tmp->yStride());
229 const int64_t _stride_pdfs_tmp_2 = int64_t(pdfs_tmp->zStride());
230 const int64_t _stride_pdfs_tmp_3 = int64_t(1 * int64_t(pdfs_tmp->fStride()));
231 const int64_t _stride_velocity_0 = int64_t(velocity->xStride());
232 const int64_t _stride_velocity_1 = int64_t(velocity->yStride());
233 const int64_t _stride_velocity_2 = int64_t(velocity->zStride());
234 const int64_t _stride_velocity_3 = int64_t(1 * int64_t(velocity->fStride()));
235 dim3 _block(uint32_c(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)), uint32_c(((1024 < ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))) ? 1024 : ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))), uint32_c(((64 < ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))) ? 64 : ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))))));
236 dim3 _grid(uint32_c(((_size_force_0 - 2) % (((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)) == 0 ? (int64_t)(_size_force_0 - 2) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)) : ((int64_t)(_size_force_0 - 2) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))) + 1)), uint32_c(((_size_force_1 - 2) % (((1024 < ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))) ? 1024 : ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))) == 0 ? (int64_t)(_size_force_1 - 2) / (int64_t)(((1024 < ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))) ? 1024 : ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))) : ((int64_t)(_size_force_1 - 2) / (int64_t)(((1024 < ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))) ? 1024 : ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) + 1)), uint32_c(((_size_force_2 - 2) % (((64 < ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))) ? 64 : ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))))) == 0 ? (int64_t)(_size_force_2 - 2) / (int64_t)(((64 < ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))) ? 64 : ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))))) : ((int64_t)(_size_force_2 - 2) / (int64_t)(((64 < ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))) ? 64 : ((_size_force_2 - 2 < ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2))))))) ? _size_force_2 - 2 : ((int64_t)(256) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2) * ((_size_force_1 - 2 < 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))) ? _size_force_1 - 2 : 2 * ((int64_t)(128) / (int64_t)(((128 < _size_force_0 - 2) ? 128 : _size_force_0 - 2)))))))))) + 1)));
237 internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda<<<_grid, _block, 0, stream>>>(_data_force, _data_pdfs, _data_pdfs_tmp, _data_velocity, _size_force_0, _size_force_1, _size_force_2, _stride_force_0, _stride_force_1, _stride_force_2, _stride_force_3, _stride_pdfs_0, _stride_pdfs_1, _stride_pdfs_2, _stride_pdfs_3, _stride_pdfs_tmp_0, _stride_pdfs_tmp_1, _stride_pdfs_tmp_2, _stride_pdfs_tmp_3, _stride_velocity_0, _stride_velocity_1, _stride_velocity_2, _stride_velocity_3);
238 pdfs->swapDataPointers(pdfs_tmp);
239}
240
241} // namespace pystencils
242} // namespace walberla
243
244#if (defined WALBERLA_CXX_COMPILER_IS_GNU) || (defined WALBERLA_CXX_COMPILER_IS_CLANG)
245#pragma GCC diagnostic pop
246#endif
247
248#if (defined WALBERLA_CXX_COMPILER_IS_INTEL)
249#pragma warning pop
250#endif
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:42
walberla::pystencils::StreamSweepSinglePrecisionCUDA::runOnCellInterval
void runOnCellInterval(const shared_ptr< StructuredBlockStorage > &blocks, const CellInterval &globalCellInterval, cell_idx_t ghostLayers, IBlock *block, gpuStream_t stream=nullptr)
Definition StreamSweepSinglePrecisionCUDA.cu:167
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:172
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_velocity_1
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_velocity_1
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_force_2
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_force_2
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_size_force_2
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const _size_force_2
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_data_pdfs_tmp
static FUNC_PREFIX float *RESTRICT const float *RESTRICT _data_pdfs_tmp
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_pdfs_tmp_0
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_pdfs_tmp_0
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_force_0
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const _stride_force_0
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_pdfs_tmp_2
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_pdfs_tmp_2
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_pdfs_3
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_pdfs_3
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_force_3
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_force_3
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_pdfs_tmp_1
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_pdfs_tmp_1
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_pdfs_tmp_3
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_pdfs_tmp_3
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_force_1
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const _stride_force_1
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_velocity_0
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_velocity_0
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_pdfs_2
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_pdfs_2
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_velocity_2
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_velocity_2
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_data_velocity
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT _data_velocity
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_size_force_1
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const _size_force_1
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::__launch_bounds__
static FUNC_PREFIX __launch_bounds__(256) void streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda(float *RESTRICT const _data_force
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_pdfs_1
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_pdfs_1
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_data_pdfs
static FUNC_PREFIX float *RESTRICT const _data_pdfs
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_size_force_0
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const _size_force_0
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla::pystencils::internal_streamsweepsingleprecisioncuda_streamsweepsingleprecisioncuda::_stride_pdfs_0
static FUNC_PREFIX float *RESTRICT const float *RESTRICT float *RESTRICT int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const int64_t const _stride_pdfs_0
Definition StreamSweepSinglePrecisionCUDA.cu:49
walberla
\file PackInfoPdfDoublePrecision.cpp \author pystencils
Definition EKWalberla.hpp:36
velocity
static Utils::Vector3d velocity(Particle const &p_ref, Particle const &p_vs)
Velocity of the virtual site.
Definition relative.cpp:64