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
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
|
/*
* Copyright (c) 2017, Miroslav Stoyanov
*
* This file is part of
* Toolkit for Adaptive Stochastic Modeling And Non-Intrusive ApproximatioN: TASMANIAN
*
* Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions
* and the following disclaimer in the documentation and/or other materials provided with the distribution.
*
* 3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse
* or promote products derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY,
* OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA,
* OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
* OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* UT-BATTELLE, LLC AND THE UNITED STATES GOVERNMENT MAKE NO REPRESENTATIONS AND DISCLAIM ALL WARRANTIES, BOTH EXPRESSED AND IMPLIED.
* THERE ARE NO EXPRESS OR IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, OR THAT THE USE OF THE SOFTWARE WILL NOT INFRINGE ANY PATENT,
* COPYRIGHT, TRADEMARK, OR OTHER PROPRIETARY RIGHTS, OR THAT THE SOFTWARE WILL ACCOMPLISH THE INTENDED RESULTS OR THAT THE SOFTWARE OR ITS USE WILL NOT RESULT IN INJURY OR DAMAGE.
* THE USER ASSUMES RESPONSIBILITY FOR ALL LIABILITIES, PENALTIES, FINES, CLAIMS, CAUSES OF ACTION, AND COSTS AND EXPENSES, CAUSED BY, RESULTING FROM OR ARISING OUT OF,
* IN WHOLE OR IN PART THE USE, STORAGE OR DISPOSAL OF THE SOFTWARE.
*/
#ifndef __TASMANIAN_SPARSE_GRID_HIP_KERNELS_HIP
#define __TASMANIAN_SPARSE_GRID_HIP_KERNELS_HIP
#include "tsgAcceleratedDataStructures.hpp"
#include "tsgDpcppBasisEvaluations.hpp"
#include "tsgDpcppWrappers.hpp"
#define _MAX_THREADS 1024
namespace TasGrid{
template<typename T>
void TasGpu::dtrans2can(AccelerationContext const *acc, bool use01, int dims, int num_x, int pad_size, double const *gpu_trans_a, double const *gpu_trans_b, T const *gpu_x_transformed, T *gpu_x_canonical){
sycl::queue *q = getSyclQueue(acc);
tasgpu_transformed_to_canonical<T, double>(q, dims, num_x, pad_size, gpu_trans_a, gpu_trans_b, gpu_x_transformed, gpu_x_canonical);
if (use01) tasgpu_m11_to_01<T>(q, dims * num_x, gpu_x_canonical);
}
template void TasGpu::dtrans2can<double>(AccelerationContext const*, bool, int, int, int, double const*, double const*, double const*, double*);
template void TasGpu::dtrans2can<float>(AccelerationContext const*, bool, int, int, int, double const*, double const*, float const*, float*);
// local polynomial basis functions, DENSE algorithm
template<typename T>
void TasGpu::devalpwpoly(AccelerationContext const *acc, int order, TypeOneDRule rule, int dims, int num_x, int num_points, const T *gpu_x, const T *gpu_nodes, const T *gpu_support, T *gpu_y){
sycl::queue *q = getSyclQueue(acc);
// order == 1 is considered "default" so that the compiler doesn't complain about missing default statement
// semilocalp cannot have order less than 2, only rule_localp can have order 0 (this gets overwrittein in makeLocalPolynomialGrid())
if (rule == rule_localp){
switch(order){
case 0:
tasgpu_devalpwpoly<T, 0, rule_localp>(q, dims, num_x, num_points, gpu_x, gpu_nodes, gpu_support, gpu_y);
break;
case 2: tasgpu_devalpwpoly<T, 2, rule_localp>(q, dims, num_x, num_points, gpu_x, gpu_nodes, gpu_support, gpu_y);
break;
default:
tasgpu_devalpwpoly<T, 1, rule_localp>(q, dims, num_x, num_points, gpu_x, gpu_nodes, gpu_support, gpu_y);
}
}else if (rule == rule_localp0){
switch(order){
case 2: tasgpu_devalpwpoly<T, 2, rule_localp0>(q, dims, num_x, num_points, gpu_x, gpu_nodes, gpu_support, gpu_y);
break;
default:
tasgpu_devalpwpoly<T, 1, rule_localp0>(q, dims, num_x, num_points, gpu_x, gpu_nodes, gpu_support, gpu_y);
}
}else if (rule == rule_localpb){
switch(order){
case 2: tasgpu_devalpwpoly<T, 2, rule_localpb>(q, dims, num_x, num_points, gpu_x, gpu_nodes, gpu_support, gpu_y);
break;
default:
tasgpu_devalpwpoly<T, 1, rule_localpb>(q, dims, num_x, num_points, gpu_x, gpu_nodes, gpu_support, gpu_y);
}
}else if (rule == rule_semilocalp){
tasgpu_devalpwpoly<T, 2, rule_semilocalp>(q, dims, num_x, num_points, gpu_x, gpu_nodes, gpu_support, gpu_y);
}else{ // rule == wavelet
tasgpu_devalpwpoly<T, 1, rule_wavelet>(q, dims, num_x, num_points, gpu_x, gpu_nodes, gpu_support, gpu_y);
}
}
template void TasGpu::devalpwpoly<double>(AccelerationContext const*, int, TypeOneDRule, int, int, int, const double*, const double*, const double*, double*);
template void TasGpu::devalpwpoly<float>(AccelerationContext const*, int, TypeOneDRule, int, int, int, const float*, const float*, const float*, float*);
// there is a switch statement that realizes templates for each combination of rule/order
// make one function that covers that switch, the rest is passed from devalpwpoly_sparse
template<typename T, int TOPLEVEL, bool fill>
inline void devalpwpoly_sparse_realize_rule_order(AccelerationContext const *acc, int order, TypeOneDRule rule, int dims, int num_x,
const T *x, const T *nodes, const T *support,
const int *hpntr, const int *hindx, int num_roots, const int *roots,
int *spntr, int *sindx, T *svals){
sycl::queue *q = getSyclQueue(acc);
if (rule == rule_localp){
switch(order){
case 0:
tasgpu_devalpwpoly_sparse<T, TOPLEVEL, 0, rule_localp, fill>
(q, dims, num_x, x, nodes, support, hpntr, hindx, num_roots, roots, spntr, sindx, svals);
break;
case 2:
tasgpu_devalpwpoly_sparse<T, TOPLEVEL, 2, rule_localp, fill>
(q, dims, num_x, x, nodes, support, hpntr, hindx, num_roots, roots, spntr, sindx, svals);
break;
default:
tasgpu_devalpwpoly_sparse<T, TOPLEVEL, 1, rule_localp, fill>
(q, dims, num_x, x, nodes, support, hpntr, hindx, num_roots, roots, spntr, sindx, svals);
}
}else if (rule == rule_localp0){
switch(order){
case 2:
tasgpu_devalpwpoly_sparse<T, TOPLEVEL, 2, rule_localp0, fill>
(q, dims, num_x, x, nodes, support, hpntr, hindx, num_roots, roots, spntr, sindx, svals);
break;
default:
tasgpu_devalpwpoly_sparse<T, TOPLEVEL, 1, rule_localp0, fill>
(q, dims, num_x, x, nodes, support, hpntr, hindx, num_roots, roots, spntr, sindx, svals);
}
}else if (rule == rule_localpb){
switch(order){
case 2:
tasgpu_devalpwpoly_sparse<T, TOPLEVEL, 2, rule_localpb, fill>
(q, dims, num_x, x, nodes, support, hpntr, hindx, num_roots, roots, spntr, sindx, svals);
break;
default:
tasgpu_devalpwpoly_sparse<T, TOPLEVEL, 1, rule_localpb, fill>
(q, dims, num_x, x, nodes, support, hpntr, hindx, num_roots, roots, spntr, sindx, svals);
}
}else{ // rule == rule_semilocalp
tasgpu_devalpwpoly_sparse<T, TOPLEVEL, 2, rule_semilocalp, fill>
(q, dims, num_x, x, nodes, support, hpntr, hindx, num_roots, roots, spntr, sindx, svals);
}
}
// local polynomial basis functions, SPARSE algorithm (2 passes, one pass to compue the non-zeros and one pass to evaluate)
template<typename T>
void TasGpu::devalpwpoly_sparse(AccelerationContext const *acc, int order, TypeOneDRule rule, int dims, int num_x, const T *gpu_x,
const GpuVector<T> &gpu_nodes, const GpuVector<T> &gpu_support,
const GpuVector<int> &gpu_hpntr, const GpuVector<int> &gpu_hindx, const GpuVector<int> &gpu_hroots,
GpuVector<int> &gpu_spntr, GpuVector<int> &gpu_sindx, GpuVector<T> &gpu_svals){
gpu_spntr.resize(acc, num_x + 1);
// call with fill == false to count the non-zeros per row of the matrix
devalpwpoly_sparse_realize_rule_order<T, 46, false>
(acc, order, rule, dims, num_x, gpu_x, gpu_nodes.data(), gpu_support.data(),
gpu_hpntr.data(), gpu_hindx.data(), (int) gpu_hroots.size(), gpu_hroots.data(), gpu_spntr.data(), 0, 0);
std::vector<int> cpu_spntr;
gpu_spntr.unload(acc, cpu_spntr);
cpu_spntr[0] = 0;
int nz = 0;
for(auto &i : cpu_spntr){
i += nz;
nz = i;
}
gpu_spntr.load(acc, cpu_spntr);
gpu_sindx.resize(acc, nz);
gpu_svals.resize(acc, nz);
// call with fill == true to load the non-zeros
devalpwpoly_sparse_realize_rule_order<T, 46, true>
(acc, order, rule, dims, num_x, gpu_x, gpu_nodes.data(), gpu_support.data(),
gpu_hpntr.data(), gpu_hindx.data(), (int) gpu_hroots.size(), gpu_hroots.data(), gpu_spntr.data(), gpu_sindx.data(), gpu_svals.data());
}
template void TasGpu::devalpwpoly_sparse<double>(AccelerationContext const*, int, TypeOneDRule, int, int,
const double*, const GpuVector<double>&, const GpuVector<double>&,
const GpuVector<int>&, const GpuVector<int>&, const GpuVector<int>&,
GpuVector<int>&, GpuVector<int>&, GpuVector<double>&);
template void TasGpu::devalpwpoly_sparse<float>(AccelerationContext const*, int, TypeOneDRule, int, int, const float*,
const GpuVector<float>&, const GpuVector<float>&,
const GpuVector<int>&, const GpuVector<int>&, const GpuVector<int>&,
GpuVector<int>&, GpuVector<int>&, GpuVector<float>&);
// Sequence Grid basis evaluations
template<typename T>
void TasGpu::devalseq(AccelerationContext const *acc, int dims, int num_x, const std::vector<int> &max_levels,
const T *gpu_x, const GpuVector<int> &num_nodes,
const GpuVector<int> &points, const GpuVector<T> &nodes, const GpuVector<T> &coeffs, T *gpu_result){
sycl::queue *q = getSyclQueue(acc);
std::vector<int> offsets(dims);
offsets[0] = 0;
for(int d=1; d<dims; d++) offsets[d] = offsets[d-1] + num_x * (max_levels[d-1] + 1);
size_t num_total = offsets[dims-1] + num_x * (max_levels[dims-1] + 1);
GpuVector<int> gpu_offsets(acc, offsets);
GpuVector<T> cache1D(acc, num_total);
tasgpu_dseq_build_cache<T>
(q, dims, num_x, gpu_x, nodes.data(), coeffs.data(), gpu_offsets.data(), num_nodes.data(), cache1D.data());
tasgpu_dseq_eval_sharedpoints<T>
(q, dims, num_x, (int) points.size() / dims, points.data(), gpu_offsets.data(), cache1D.data(), gpu_result);
}
template void TasGpu::devalseq<double>(AccelerationContext const*, int dims, int num_x, const std::vector<int> &max_levels,
const double *gpu_x, const GpuVector<int> &num_nodes,
const GpuVector<int> &points, const GpuVector<double> &nodes, const GpuVector<double> &coeffs, double *gpu_result);
template void TasGpu::devalseq<float>(AccelerationContext const*, int dims, int num_x, const std::vector<int> &max_levels,
const float *gpu_x, const GpuVector<int> &num_nodes,
const GpuVector<int> &points, const GpuVector<float> &nodes, const GpuVector<float> &coeffs, float *gpu_result);
// Fourier Grid basis evaluations
template<typename T>
void TasGpu::devalfor(AccelerationContext const *acc, int dims, int num_x, const std::vector<int> &max_levels, const T *gpu_x,
const GpuVector<int> &num_nodes, const GpuVector<int> &points, T *gpu_wreal, typename GpuVector<T>::value_type *gpu_wimag){
sycl::queue *q = getSyclQueue(acc);
std::vector<int> max_nodes(dims);
for(int j=0; j<dims; j++){
int n = 1;
for(int i=0; i<max_levels[j]; i++) n *= 3;
max_nodes[j] = n;
}
std::vector<int> offsets(dims);
offsets[0] = 0;
for(int d=1; d<dims; d++) offsets[d] = offsets[d-1] + 2 * num_x * (max_nodes[d-1] + 1);
size_t num_total = offsets[dims-1] + 2 * num_x * (max_nodes[dims-1] + 1);
GpuVector<int> gpu_offsets(acc, offsets);
GpuVector<T> cache1D(acc, num_total);
tasgpu_dfor_build_cache<T>(q, dims, num_x, gpu_x, gpu_offsets.data(), num_nodes.data(), cache1D.data());
if (gpu_wimag == 0){
tasgpu_dfor_eval_sharedpoints<T, true>
(q, dims, num_x, (int) points.size() / dims, points.data(), gpu_offsets.data(), cache1D.data(), gpu_wreal, 0);
}else{
tasgpu_dfor_eval_sharedpoints<T, false>
(q, dims, num_x, (int) points.size() / dims, points.data(), gpu_offsets.data(), cache1D.data(), gpu_wreal, gpu_wimag);
}
}
template void TasGpu::devalfor<double>(AccelerationContext const*, int, int, const std::vector<int>&, const double*, const GpuVector<int>&, const GpuVector<int>&, double*, double*);
template void TasGpu::devalfor<float>(AccelerationContext const*, int, int, const std::vector<int>&, const float*, const GpuVector<int>&, const GpuVector<int>&, float*, float*);
template<typename T>
void TasGpu::devalglo(AccelerationContext const *acc, bool is_nested, bool is_clenshawcurtis0, int dims, int num_x, int num_p, int num_basis,
T const *gpu_x, GpuVector<T> const &nodes, GpuVector<T> const &coeff, GpuVector<T> const &tensor_weights,
GpuVector<int> const &nodes_per_level, GpuVector<int> const &offset_per_level, GpuVector<int> const &map_dimension, GpuVector<int> const &map_level,
GpuVector<int> const &active_tensors, GpuVector<int> const &active_num_points, GpuVector<int> const &dim_offsets,
GpuVector<int> const &map_tensor, GpuVector<int> const &map_index, GpuVector<int> const &map_reference, T *gpu_result){
GpuVector<T> cache(acc, num_x, num_basis);
sycl::queue *q = getSyclQueue(acc);
if (is_nested){
if (is_clenshawcurtis0){
tasgpu_dglo_build_cache<T, true, true>
(q, dims, num_x, (int) map_dimension.size(), gpu_x, nodes.data(), coeff.data(),
nodes_per_level.data(), offset_per_level.data(), dim_offsets.data(),
map_dimension.data(), map_level.data(), cache.data());
}else{
tasgpu_dglo_build_cache<T, true, false>
(q, dims, num_x, (int) map_dimension.size(), gpu_x, nodes.data(), coeff.data(),
nodes_per_level.data(), offset_per_level.data(), dim_offsets.data(),
map_dimension.data(), map_level.data(), cache.data());
}
}else{
tasgpu_dglo_build_cache<T, false, false>
(q, dims, num_x, (int) map_dimension.size(), gpu_x, nodes.data(), coeff.data(),
nodes_per_level.data(), offset_per_level.data(), dim_offsets.data(),
map_dimension.data(), map_level.data(), cache.data());
}
tasgpu_dglo_eval_zero<T>(q, Utils::size_mult(num_x, num_p), gpu_result);
tasgpu_dglo_eval_sharedpoints<T>
(q, dims, num_x, (int) map_tensor.size(), num_p, cache.data(),
tensor_weights.data(), offset_per_level.data(), dim_offsets.data(), active_tensors.data(), active_num_points.data(),
map_tensor.data(), map_index.data(), map_reference.data(), gpu_result);
}
template void TasGpu::devalglo<double>(AccelerationContext const*, bool, bool, int, int, int, int,
double const*, GpuVector<double> const&, GpuVector<double> const&, GpuVector<double> const&,
GpuVector<int> const&, GpuVector<int> const&, GpuVector<int> const&, GpuVector<int> const&,
GpuVector<int> const&, GpuVector<int> const&, GpuVector<int> const&,
GpuVector<int> const&, GpuVector<int> const&, GpuVector<int> const&, double*);
template void TasGpu::devalglo<float>(AccelerationContext const*, bool, bool, int, int, int, int,
float const*, GpuVector<float> const&, GpuVector<float> const&, GpuVector<float> const&,
GpuVector<int> const&, GpuVector<int> const&, GpuVector<int> const&, GpuVector<int> const&,
GpuVector<int> const&, GpuVector<int> const&, GpuVector<int> const&,
GpuVector<int> const&, GpuVector<int> const&, GpuVector<int> const&, float*);
void TasGpu::fillDataGPU(AccelerationContext const *acc, double value, long long n, long long stride, double data[]){
sycl::queue *q = getSyclQueue(acc);
if (stride == 1){
q->submit([&](sycl::handler& h) {
h.parallel_for<class tasgpu_vfill_kernel>(sycl::range<1>{static_cast<size_t>(n),}, [=](sycl::id<1> threadId){
data[threadId[0]] = value;
});
});
q->wait();
}else{
q->submit([&](sycl::handler& h) {
h.parallel_for<class tasgpu_sfill_kernel>(sycl::range<1>{static_cast<size_t>(n),}, [=](sycl::id<1> threadId){
data[threadId[0] * stride] = value;
});
});
q->wait();
}
}
}
#endif
|
