/usr/include/trilinos/KokkosKernels_SPGEMMHandle.hpp is in libtrilinos-kokkos-kernels-dev 12.12.1-5.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 | /*
//@HEADER
// ************************************************************************
//
// KokkosKernels 0.9: Linear Algebra and Graph Kernels
// Copyright 2017 Sandia Corporation
//
// Under the terms of Contract DE-AC04-94AL85000 with Sandia Corporation,
// the U.S. Government retains certain rights in this software.
//
// 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 Corporation nor the names of the
// contributors may be used to endorse or promote products derived from
// this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY SANDIA CORPORATION "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 SANDIA CORPORATION OR THE
// 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.
//
// Questions? Contact Siva Rajamanickam (srajama@sandia.gov)
//
// ************************************************************************
//@HEADER
*/
#include <Kokkos_MemoryTraits.hpp>
#include <Kokkos_Core.hpp>
#include <iostream>
//#define KERNELS_HAVE_CUSPARSE
#ifdef KERNELS_HAVE_CUSPARSE
#include "cusparse.h"
#endif
#ifndef _SPGEMMHANDLE_HPP
#define _SPGEMMHANDLE_HPP
//#define VERBOSE
namespace KokkosKernels{
namespace Experimental{
namespace Graph{
//TODO:SPGEMM_KK_MEMORY2 option is for testing in openmp.
//it wont work on cuda, not bind to a test program.
//hidden parameter for StringToSPGEMMAlgorithm for now.
enum SPGEMMAlgorithm{SPGEMM_DEFAULT, SPGEMM_DEBUG, SPGEMM_SERIAL,
SPGEMM_CUSPARSE, SPGEMM_CUSP, SPGEMM_MKL, SPGEMM_MKL2PHASE, SPGEMM_VIENNA,
//SPGEMM_KK1, SPGEMM_KK2, SPGEMM_KK3, SPGEMM_KK4,
SPGEMM_KK_MULTIMEM,
SPGEMM_KK_SPEED, SPGEMM_KK_MEMORY, SPGEMM_KK_MEMORY2, SPGEMM_KK_COLOR, SPGEMM_KK_MULTICOLOR, SPGEMM_KK_MULTICOLOR2, SPGEMM_KK_MEMSPEED};
template <class lno_row_view_t_,
class lno_nnz_view_t_,
class scalar_nnz_view_t_,
class ExecutionSpace,
class TemporaryMemorySpace,
class PersistentMemorySpace>
class SPGEMMHandle{
public:
typedef ExecutionSpace HandleExecSpace;
typedef TemporaryMemorySpace HandleTempMemorySpace;
typedef PersistentMemorySpace HandlePersistentMemorySpace;
typedef lno_row_view_t_ in_lno_row_view_t;
typedef lno_nnz_view_t_ in_lno_nnz_view_t;
typedef scalar_nnz_view_t_ in_scalar_nnz_view_t;
typedef typename in_lno_row_view_t::non_const_value_type size_type;
/*
typedef typename in_lno_row_view_t::array_layout row_lno_view_array_layout;
typedef typename in_lno_row_view_t::device_type row_lno_view_device_t;
typedef typename in_lno_row_view_t::memory_traits row_lno_view_memory_traits;
typedef typename in_lno_row_view_t::HostMirror row_lno_host_view_t; //Host view type
*/
//typedef typename idx_memory_traits::MemorySpace MyMemorySpace;
typedef typename in_lno_nnz_view_t::non_const_value_type nnz_lno_t;
/*
typedef typename in_lno_nnz_view_t::array_layout nnz_lno_view_array_layout;
typedef typename in_lno_nnz_view_t::device_type nnz_lno_view_device_t;
typedef typename in_lno_nnz_view_t::memory_traits nnz_lno_view_memory_traits;
typedef typename in_lno_nnz_view_t::HostMirror nnz_lno_host_view_t; //Host view type
*/
//typedef typename idx_edge_memory_traits::MemorySpace MyEdgeMemorySpace;
typedef typename in_scalar_nnz_view_t::non_const_value_type nnz_scalar_t;
/*
typedef typename in_scalar_nnz_view_t::array_layout nnz_scalar_view_array_layout;
typedef typename in_scalar_nnz_view_t::device_type nnz_scalar_view_device_t;
typedef typename in_scalar_nnz_view_t::memory_traits nnz_scalar_view_memory_traits;
typedef typename in_scalar_nnz_view_t::HostMirror nnz_scalar_view_t; //Host view type
*/
/*
typedef typename in_lno_row_view_t::const_data_type const_row_lno_t;
typedef typename in_lno_row_view_t::non_const_data_type non_const_row_lno_t;
*/
typedef typename in_lno_row_view_t::const_type const_lno_row_view_t;
typedef typename in_lno_row_view_t::non_const_type non_const_lno_row_view_t;
/*
typedef typename in_lno_nnz_view_t::const_data_type const_nnz_lno_t;
typedef typename in_lno_nnz_view_t::non_const_data_type non_const_nnz_lno_t;
*/
typedef typename in_lno_nnz_view_t::const_type const_lno_nnz_view_t;
typedef typename in_lno_nnz_view_t::non_const_type non_const_lno_nnz_view_t;
/*
typedef typename in_scalar_nnz_view_t::const_data_type const_nnz_scalar_t;
typedef typename in_scalar_nnz_view_t::non_const_data_type non_const_nnz_scalar_t;
*/
typedef typename in_scalar_nnz_view_t::const_type const_scalar_nnz_view_t;
typedef typename in_scalar_nnz_view_t::non_const_type non_const_scalar_nnz_view_t;
typedef typename Kokkos::View<size_type *, HandleTempMemorySpace> row_lno_temp_work_view_t;
typedef typename Kokkos::View<size_type *, HandlePersistentMemorySpace> row_lno_persistent_work_view_t;
typedef typename row_lno_persistent_work_view_t::HostMirror row_lno_persistent_work_host_view_t; //Host view type
typedef typename Kokkos::View<nnz_scalar_t *, HandleTempMemorySpace> scalar_temp_work_view_t;
typedef typename Kokkos::View<nnz_scalar_t *, HandlePersistentMemorySpace> scalar_persistent_work_view_t;
typedef typename Kokkos::View<nnz_lno_t *, HandleTempMemorySpace> nnz_lno_temp_work_view_t;
typedef typename Kokkos::View<nnz_lno_t *, HandlePersistentMemorySpace> nnz_lno_persistent_work_view_t;
typedef typename nnz_lno_persistent_work_view_t::HostMirror nnz_lno_persistent_work_host_view_t; //Host view type
#ifdef KERNELS_HAVE_CUSPARSE
struct cuSparseHandleType{
cusparseHandle_t handle;
cusparseOperation_t transA;
cusparseOperation_t transB;
cusparseMatDescr_t a_descr;
cusparseMatDescr_t b_descr;
cusparseMatDescr_t c_descr;
cuSparseHandleType(bool transposeA, bool transposeB){
cusparseStatus_t status;
status= cusparseCreate(&handle);
if (status != CUSPARSE_STATUS_SUCCESS) {
throw std::runtime_error ("cusparseCreate ERROR\n");
return;
}
cusparseSetPointerMode(handle, CUSPARSE_POINTER_MODE_HOST);
if (transposeA){
transA = CUSPARSE_OPERATION_TRANSPOSE;
}
else {
transA = CUSPARSE_OPERATION_NON_TRANSPOSE;
}
if (transposeB){
transB = CUSPARSE_OPERATION_TRANSPOSE;
}
else {
transB = CUSPARSE_OPERATION_NON_TRANSPOSE;
}
status = cusparseCreateMatDescr(&a_descr);
if (status != CUSPARSE_STATUS_SUCCESS) {
throw std::runtime_error ("cusparseCreateMatDescr a_descr ERROR\n");
return;
}
cusparseSetMatType(a_descr,CUSPARSE_MATRIX_TYPE_GENERAL);
cusparseSetMatIndexBase(a_descr,CUSPARSE_INDEX_BASE_ZERO);
status = cusparseCreateMatDescr(&b_descr);
if (status != CUSPARSE_STATUS_SUCCESS) {
throw std::runtime_error ("cusparseCreateMatDescr b_descr ERROR\n");
return;
}
cusparseSetMatType(b_descr,CUSPARSE_MATRIX_TYPE_GENERAL);
cusparseSetMatIndexBase(b_descr,CUSPARSE_INDEX_BASE_ZERO);
status = cusparseCreateMatDescr(&c_descr);
if (status != CUSPARSE_STATUS_SUCCESS) {
throw std::runtime_error ("cusparseCreateMatDescr c_descr ERROR\n");
return;
}
cusparseSetMatType(c_descr,CUSPARSE_MATRIX_TYPE_GENERAL);
cusparseSetMatIndexBase(c_descr,CUSPARSE_INDEX_BASE_ZERO);
}
~cuSparseHandleType(){
cusparseDestroyMatDescr(a_descr);
cusparseDestroyMatDescr(b_descr);
cusparseDestroyMatDescr(c_descr);
cusparseDestroy(handle);
}
};
typedef cuSparseHandleType SPGEMMcuSparseHandleType;
#endif
private:
SPGEMMAlgorithm algorithm_type;
size_type result_nnz_size;
bool called_symbolic;
bool called_numeric;
int suggested_vector_size;
int suggested_team_size;
nnz_lno_t max_nnz_inresult;
nnz_lno_t max_nnz_compressed_result;
row_lno_temp_work_view_t compressed_b_rowmap;// compressed_b_set_begins, compressed_b_set_nexts;
nnz_lno_temp_work_view_t compressed_b_set_indices, compressed_b_sets;
row_lno_temp_work_view_t compressed_c_rowmap;
nnz_lno_temp_work_view_t c_column_indices;
row_lno_temp_work_view_t tranpose_a_xadj, tranpose_b_xadj, tranpose_c_xadj;
nnz_lno_temp_work_view_t tranpose_a_adj, tranpose_b_adj, tranpose_c_adj;
bool transpose_a,transpose_b, transpose_c_symbolic;
nnz_lno_t num_colors;
nnz_lno_persistent_work_host_view_t color_xadj;
nnz_lno_persistent_work_view_t color_adj, vertex_colors;
nnz_lno_t num_multi_colors, num_used_colors;
double multi_color_scale;
int mkl_sort_option;
#ifdef KERNELS_HAVE_CUSPARSE
SPGEMMcuSparseHandleType *cuSPARSEHandle;
#endif
public:
typename Kokkos::View<int *, HandlePersistentMemorySpace> persistent_c_xadj, persistent_a_xadj, persistent_b_xadj, persistent_a_adj, persistent_b_adj;
bool mkl_keep_output;
bool mkl_convert_to_1base;
void set_mkl_sort_option(int mkl_sort_option_){
this->mkl_sort_option = mkl_sort_option_;
}
int get_mkl_sort_option(){
return this->mkl_sort_option;
}
void set_c_column_indices(nnz_lno_temp_work_view_t c_col_indices_){
this->c_column_indices = c_col_indices_;
}
nnz_lno_temp_work_view_t get_c_column_indices(){
return this->c_column_indices;
}
void set_color_xadj(
nnz_lno_t num_colors_,
nnz_lno_persistent_work_host_view_t color_xadj_,
nnz_lno_persistent_work_view_t color_adj_,
nnz_lno_persistent_work_view_t vertex_colors_,
nnz_lno_t num_multi_colors_, nnz_lno_t num_used_colors_){
num_colors = num_colors_;
color_xadj = color_xadj_;
color_adj = color_adj_;
vertex_colors = vertex_colors_;
num_multi_colors = num_multi_colors_;
num_used_colors = num_used_colors_;
}
/**
* \brief sets the result nnz size.
* \param result_nnz_size: size of the output matrix.
*/
void set_c_nnz(size_type result_nnz_size_){
this->result_nnz_size = result_nnz_size_;
}
/**
* \brief returns the result nnz size.
*/
size_type get_c_nnz(){
return this->result_nnz_size;
}
void set_multi_color_scale(double multi_color_scale_){
this->multi_color_scale = multi_color_scale_;
}
double get_multi_color_scale(){
return this->multi_color_scale;
}
void get_color_xadj(
nnz_lno_t &num_colors_,
nnz_lno_persistent_work_host_view_t &color_xadj_,
nnz_lno_persistent_work_view_t &color_adj_,
nnz_lno_persistent_work_view_t &vertex_colors_,
nnz_lno_t &num_multi_colors_, nnz_lno_t &num_used_colors_){
num_colors_ = num_colors;
color_xadj_ = color_xadj;
color_adj_ = color_adj ;
num_multi_colors_ = num_multi_colors;
num_used_colors_ = num_used_colors ;
vertex_colors_ = vertex_colors;
}
void set_compressed_c(
row_lno_temp_work_view_t compressed_c_rowmap_){
compressed_c_rowmap = compressed_c_rowmap_;
}
void get_compressed_c(
row_lno_temp_work_view_t &compressed_c_rowmap_){
compressed_c_rowmap_ = compressed_c_rowmap;
}
//TODO: store transpose here.
void get_c_transpose_symbolic(){}
void get_compressed_b(
row_lno_temp_work_view_t &compressed_b_rowmap_,
nnz_lno_temp_work_view_t &compressed_b_set_indices_,
nnz_lno_temp_work_view_t &compressed_b_sets_,
row_lno_temp_work_view_t &compressed_b_set_begins_,
row_lno_temp_work_view_t &compressed_b_set_nexts_){
compressed_b_rowmap_ = compressed_b_rowmap;
compressed_b_set_indices_ = compressed_b_set_indices;
compressed_b_sets_ = compressed_b_sets;
}
/**
* \brief Default constructor.
*/
SPGEMMHandle(SPGEMMAlgorithm gs = SPGEMM_DEFAULT):
algorithm_type(gs), result_nnz_size(0),
called_symbolic(false), called_numeric(false),
suggested_vector_size(0), suggested_team_size(0), max_nnz_inresult(0),
c_column_indices(),
tranpose_a_xadj(), tranpose_b_xadj(), tranpose_c_xadj(),
tranpose_a_adj(), tranpose_b_adj(), tranpose_c_adj(),
transpose_a(false),transpose_b(false), transpose_c_symbolic(false),
num_colors(0),
color_xadj(), color_adj(), vertex_colors(), num_multi_colors(0),num_used_colors(0), multi_color_scale(1), mkl_sort_option(7),
persistent_a_xadj(), persistent_b_xadj(), persistent_a_adj(), persistent_b_adj(),
mkl_keep_output(true),
mkl_convert_to_1base(true)
#ifdef KERNELS_HAVE_CUSPARSE
,cuSPARSEHandle(NULL)
#endif
{
if (gs == SPGEMM_DEFAULT){
this->choose_default_algorithm();
}
}
virtual ~SPGEMMHandle(){
#ifdef KERNELS_HAVE_CUSgPARSE
this->destroy_cuSPARSE_Handle();
#endif
};
#ifdef KERNELS_HAVE_CUSPARSE
void create_cuSPARSE_Handle(bool transA, bool transB){
this->destroy_cuSPARSE_Handle();
this->cuSPARSEHandle = new cuSparseHandleType(transA, transB);
}
void destroy_cuSPARSE_Handle(){
if (this->cuSPARSEHandle != NULL){
delete this->cuSPARSEHandle;
this->cuSPARSEHandle = NULL;
}
}
SPGEMMcuSparseHandleType *get_cuSparseHandle(){
return this->cuSPARSEHandle;
}
#endif
/** \brief Chooses best algorithm based on the execution space. COLORING_EB if cuda, COLORING_VB otherwise.
*/
void choose_default_algorithm(){
#if defined( KOKKOS_HAVE_SERIAL )
if (Kokkos::Impl::is_same< Kokkos::Serial , ExecutionSpace >::value){
this->algorithm_type = SPGEMM_SERIAL;
#ifdef VERBOSE
std::cout << "Serial Execution Space, Default Algorithm: SPGEMM_SERIAL" << std::endl;
#endif
}
#endif
#if defined( KOKKOS_HAVE_PTHREAD )
if (Kokkos::Impl::is_same< Kokkos::Threads , ExecutionSpace >::value){
this->algorithm_type = SPGEMM_SERIAL;
#ifdef VERBOSE
std::cout << "PTHREAD Execution Space, Default Algorithm: SPGEMM_SERIAL" << std::endl;
#endif
}
#endif
#if defined( KOKKOS_HAVE_OPENMP )
if (Kokkos::Impl::is_same< Kokkos::OpenMP, ExecutionSpace >::value){
this->algorithm_type = SPGEMM_SERIAL;
#ifdef VERBOSE
std::cout << "OpenMP Execution Space, Default Algorithm: SPGEMM_SERIAL" << std::endl;
#endif
}
#endif
#if defined( KOKKOS_HAVE_CUDA )
if (Kokkos::Impl::is_same<Kokkos::Cuda, ExecutionSpace >::value){
this->algorithm_type = SPGEMM_CUSPARSE;
#ifdef VERBOSE
std::cout << "Cuda Execution Space, Default Algorithm: SPGEMM_CUSPARSE" << std::endl;
#endif
}
#endif
#if defined( KOKKOS_HAVE_QTHREAD)
if (Kokkos::Impl::is_same< Kokkos::Qthread, ExecutionSpace >::value){
this->algorithm_type = SPGEMM_SERIAL;
#ifdef VERBOSE
std::cout << "Qthread Execution Space, Default Algorithm: SPGEMM_SERIAL" << std::endl;
#endif
}
#endif
}
//getters
SPGEMMAlgorithm get_algorithm_type() const {return this->algorithm_type;}
bool is_symbolic_called(){return this->called_symbolic;}
bool is_numeric_called(){return this->called_numeric;}
nnz_lno_t get_max_result_nnz() const{
return this->max_nnz_inresult ;
}
nnz_lno_t get_max_compresed_result_nnz() const{
return this->max_nnz_compressed_result ;
}
//setters
void set_algorithm_type(const SPGEMMAlgorithm &sgs_algo){this->algorithm_type = sgs_algo;}
void set_call_symbolic(bool call = true){this->called_symbolic = call;}
void set_call_numeric(bool call = true){this->called_numeric = call;}
void set_max_result_nnz(nnz_lno_t num_result_nnz_){
this->max_nnz_inresult = num_result_nnz_;
}
void set_max_compresed_result_nnz(nnz_lno_t num_result_nnz_){
this->max_nnz_compressed_result = num_result_nnz_;
}
void vector_team_size(
int max_allowed_team_size,
int &suggested_vector_size_,
int &suggested_team_size_,
size_type nr, size_type nnz){
//suggested_team_size_ = this->suggested_team_size = 1;
//suggested_vector_size_=this->suggested_vector_size = 1;
//return;
if (this->suggested_team_size && this->suggested_vector_size) {
suggested_vector_size_ = this->suggested_vector_size;
suggested_team_size_ = this->suggested_team_size;
return;
}
#if defined( KOKKOS_HAVE_SERIAL )
if (Kokkos::Impl::is_same< Kokkos::Serial , ExecutionSpace >::value){
suggested_vector_size_ = this->suggested_vector_size = 1;
suggested_team_size_ = this->suggested_team_size = max_allowed_team_size;
return;
}
#endif
#if defined( KOKKOS_HAVE_PTHREAD )
if (Kokkos::Impl::is_same< Kokkos::Threads , ExecutionSpace >::value){
suggested_vector_size_ = this->suggested_vector_size = 1;
suggested_team_size_ = this->suggested_team_size = max_allowed_team_size;
return;
}
#endif
#if defined( KOKKOS_HAVE_OPENMP )
if (Kokkos::Impl::is_same< Kokkos::OpenMP, ExecutionSpace >::value){
suggested_vector_size_ = this->suggested_vector_size = 1;
suggested_team_size_ = this->suggested_team_size = max_allowed_team_size;
}
#endif
#if defined( KOKKOS_HAVE_CUDA )
if (Kokkos::Impl::is_same<Kokkos::Cuda, ExecutionSpace >::value){
this->suggested_vector_size = nnz / double (nr) + 0.5;
if (this->suggested_vector_size <= 3){
this->suggested_vector_size = 2;
}
else if (this->suggested_vector_size <= 6){
this->suggested_vector_size = 4;
}
else if (this->suggested_vector_size <= 12){
this->suggested_vector_size = 8;
}
else if (this->suggested_vector_size <= 24){
this->suggested_vector_size = 16;
}
else {
this->suggested_vector_size = 32;
}
suggested_vector_size_ = this->suggested_vector_size;
this->suggested_team_size= suggested_team_size_ = max_allowed_team_size / this->suggested_vector_size;
}
#endif
#if defined( KOKKOS_HAVE_QTHREAD)
if (Kokkos::Impl::is_same< Kokkos::Qthread, ExecutionSpace >::value){
suggested_vector_size_ = this->suggested_vector_size = 1;
suggested_team_size_ = this->suggested_team_size = max_allowed_team_size;
}
#endif
}
};
inline SPGEMMAlgorithm StringToSPGEMMAlgorithm(std::string & name) {
if(name=="SPGEMM_DEFAULT") return SPGEMM_DEFAULT;
else if(name=="SPGEMM_DEBUG") return SPGEMM_DEBUG;
else if(name=="SPGEMM_SERIAL") return SPGEMM_SERIAL;
else if(name=="SPGEMM_CUSPARSE") return SPGEMM_CUSPARSE;
else if(name=="SPGEMM_CUSP") return SPGEMM_CUSP;
else if(name=="SPGEMM_MKL") return SPGEMM_MKL;
else if(name=="SPGEMM_VIENNA") return SPGEMM_VIENNA;
else if(name=="SPGEMM_KK_SPEED") return SPGEMM_KK_SPEED;
else if(name=="SPGEMM_KK_MEMORY") return SPGEMM_KK_MEMORY;
else if(name=="SPGEMM_KK_COLOR") return SPGEMM_KK_COLOR;
else if(name=="SPGEMM_KK_MULTICOLOR") return SPGEMM_KK_MULTICOLOR;
else if(name=="SPGEMM_KK_MULTICOLOR2") return SPGEMM_KK_MULTICOLOR2;
else if(name=="SPGEMM_KK_MEMSPEED") return SPGEMM_KK_MEMSPEED;
else
throw std::runtime_error("Invalid SPGEMMAlgorithm name");
}
}
}
}
#endif
|