/usr/include/trilinos/Stokhos_CooProductTensor.hpp is in libtrilinos-stokhos-dev 12.10.1-3.
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 | // @HEADER
// ***********************************************************************
//
// Stokhos Package
// Copyright (2009) Sandia Corporation
//
// Under terms of Contract DE-AC04-94AL85000, there is a non-exclusive
// license for use of this work by or on behalf of the U.S. Government.
//
// 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 Eric T. Phipps (etphipp@sandia.gov).
//
// ***********************************************************************
// @HEADER
#ifndef STOKHOS_COO_PRODUCT_TENSOR_HPP
#define STOKHOS_COO_PRODUCT_TENSOR_HPP
#include "Kokkos_Core.hpp"
#include "Stokhos_Multiply.hpp"
#include "Stokhos_ProductBasis.hpp"
#include "Stokhos_Sparse3Tensor.hpp"
#include "Teuchos_ParameterList.hpp"
#include "Kokkos_Core.hpp"
//----------------------------------------------------------------------------
//----------------------------------------------------------------------------
namespace Stokhos {
/** \brief Sparse product tensor using 'COO'-like storage format.
*
* Entries are sorted by stochastic 'row', requiring segmented-reduction
* based on row index.
*
* Third template argument determines whether to store (i,j,k) triple in a
* single value.
*/
template< typename ValueType, class ExecutionSpace, bool PackIndex >
class CooProductTensor {};
/** \brief Specialization of CooProductTensor for packed (i,j,k)
*/
template< typename ValueType, class ExecutionSpace >
class CooProductTensor<ValueType,ExecutionSpace,true> {
public:
typedef ExecutionSpace execution_space;
typedef typename execution_space::size_type size_type;
typedef ValueType value_type;
private:
// Number of bits available for each index
static const size_type bits = (sizeof(size_type)*8) / 3;
// Mask for packing index
static const size_type mask = (1 << bits)-1;
typedef Kokkos::View< value_type[], execution_space > vec_type;
typedef Kokkos::View< size_type[], execution_space > coord_array_type;
typedef Kokkos::View< value_type[], execution_space > value_array_type;
coord_array_type m_coord;
value_array_type m_value;
size_type m_dim;
size_type m_flops;
// Pack (i,j,k) into a single integer
static size_type
pack( const size_type i, const size_type j, const size_type k ) {
const size_type ii = i & mask;
const size_type jj = j & mask;
const size_type kk = k & mask;
size_type ijk = ii | (jj << bits) | (kk << 2*bits);
return ijk;
}
KOKKOS_INLINE_FUNCTION
void unpack( size_type ijk, size_type& i, size_type& j, size_type& k ) const {
i = ijk & mask; ijk >>= bits;
j = ijk & mask;
k = ijk >> bits;
}
public:
//! Maximum index storable by packed approach
static const size_type max_index = 1 << bits;
inline
~CooProductTensor() {}
inline
CooProductTensor() :
m_coord(),
m_value(),
m_dim(0),
m_flops(0) {}
inline
CooProductTensor( const CooProductTensor & rhs ) :
m_coord( rhs.m_coord ),
m_value( rhs.m_value ),
m_dim( rhs.m_dim ),
m_flops( rhs.m_flops ) {}
inline
CooProductTensor & operator = ( const CooProductTensor & rhs )
{
m_coord = rhs.m_coord;
m_value = rhs.m_value;
m_dim = rhs.m_dim;
m_flops = rhs.m_flops;
return *this;
}
/** \brief Dimension of the tensor. */
KOKKOS_INLINE_FUNCTION
size_type dimension() const { return m_dim; }
/** \brief Number of sparse entries. */
KOKKOS_INLINE_FUNCTION
size_type entry_count() const { return m_coord.dimension_0(); }
/** \brief Get (i,j,k) coordinates of an entry */
KOKKOS_INLINE_FUNCTION
void coord( const size_type entry,
size_type& i, size_type& j, size_type& k ) const {
unpack(m_coord(entry), i, j, k);
}
/** \brief Value of an entry */
KOKKOS_INLINE_FUNCTION
const value_type & value( const size_type entry ) const
{ return m_value(entry); }
/** \brief Number of non-zero's */
KOKKOS_INLINE_FUNCTION
size_type num_non_zeros() const { return m_coord.dimension_0(); }
/** \brief Number flop's per multiply-add */
KOKKOS_INLINE_FUNCTION
size_type num_flops() const { return m_flops; }
template <typename OrdinalType>
static CooProductTensor
create( const Stokhos::ProductBasis<OrdinalType,ValueType>& basis,
const Stokhos::Sparse3Tensor<OrdinalType,ValueType>& Cijk,
const Teuchos::ParameterList& params = Teuchos::ParameterList())
{
typedef Stokhos::Sparse3Tensor<OrdinalType,ValueType> Cijk_type;
typedef typename Cijk_type::i_iterator i_iterator;
typedef typename Cijk_type::ik_iterator k_iterator;
typedef typename Cijk_type::ikj_iterator j_iterator;
// Compute entry count
size_type entry_count = 0;
for (i_iterator i_it = Cijk.i_begin(); i_it!=Cijk.i_end(); ++i_it) {
for (k_iterator k_it = Cijk.k_begin(i_it); k_it != Cijk.k_end(i_it);
++k_it) {
OrdinalType k = index(k_it);
for (j_iterator j_it = Cijk.j_begin(k_it); j_it != Cijk.j_end(k_it);
++j_it) {
OrdinalType j = index(j_it);
if (j >= k) {
++entry_count;
}
}
}
}
// Align entry_count
#if defined( KOKKOS_HAVE_CUDA )
enum { Align = Kokkos::Impl::is_same<ExecutionSpace,Kokkos::Cuda>::value ? 32 : 1 };
#else
enum { Align = 1 };
#endif
entry_count = (entry_count+Align-1) & ~(Align-1);
TEUCHOS_ASSERT(entry_count % Align == 0);
// Allocate tensor data
CooProductTensor tensor;
tensor.m_coord = coord_array_type( "tensor_coord", entry_count );
tensor.m_value = value_array_type( "tensor_value", entry_count );
tensor.m_dim = basis.size();
tensor.m_flops = 5*entry_count + tensor.m_dim;
// Create mirror, is a view if is host memory
typename coord_array_type::HostMirror
host_coord = Kokkos::create_mirror_view( tensor.m_coord );
typename value_array_type::HostMirror
host_value = Kokkos::create_mirror_view( tensor.m_value );
size_type n = 0;
OrdinalType i=0, j=0, k=0;
for (i_iterator i_it = Cijk.i_begin(); i_it!=Cijk.i_end(); ++i_it) {
i = index(i_it);
for (k_iterator k_it = Cijk.k_begin(i_it); k_it != Cijk.k_end(i_it);
++k_it) {
k = index(k_it);
for (j_iterator j_it = Cijk.j_begin(k_it); j_it != Cijk.j_end(k_it);
++j_it) {
j = index(j_it);
ValueType c = Stokhos::value(j_it);
if (j >= k) {
host_value(n) = (j != k) ? c : 0.5*c;
host_coord(n) = pack(i,j,k);
++n;
}
}
}
}
for (; n < entry_count; ++n) {
host_value(n) = 0.0;
host_coord(n) = pack(i,j,k);
}
// Copy data to device if necessary
Kokkos::deep_copy( tensor.m_coord, host_coord );
Kokkos::deep_copy( tensor.m_value, host_value );
return tensor;
}
void print(std::ostream& os) const {
size_type num_entry = entry_count();
typename coord_array_type::HostMirror
host_coord = Kokkos::create_mirror_view( m_coord );
typename value_array_type::HostMirror
host_value = Kokkos::create_mirror_view( m_value );
Kokkos::deep_copy( host_coord, m_coord );
Kokkos::deep_copy( host_value, m_value );
os << "CooProductTensor: dim = "
<< dimension() << ", entry_count = "
<< num_entry << std::endl
<< "Entries: i j k : cijk" << std::endl;
for (size_type l=0; l<num_entry; ++l) {
size_type i,j,k;
unpack(host_coord(l), i, j, k);
ValueType cijk = host_value(l);
os << "\t " << l << " : " << i << " " << j << " " << k << " = " << cijk
<< std::endl;
if (l > 0 && l % 32 == 0) os << std::endl;
}
}
};
/** \brief Specialization of CooProductTensor for unpacked (i,j,k)
*/
template< typename ValueType, class ExecutionSpace>
class CooProductTensor<ValueType,ExecutionSpace,false> {
public:
typedef ExecutionSpace execution_space;
typedef typename execution_space::size_type size_type;
typedef ValueType value_type;
private:
typedef Kokkos::View< value_type[], execution_space > vec_type;
typedef Kokkos::View< size_type[][3], execution_space > coord_array_type;
typedef Kokkos::View< value_type[], execution_space > value_array_type;
coord_array_type m_coord;
value_array_type m_value;
size_type m_dim;
size_type m_flops;
public:
inline
~CooProductTensor() {}
inline
CooProductTensor() :
m_coord(),
m_value(),
m_dim(0),
m_flops(0) {}
inline
CooProductTensor( const CooProductTensor & rhs ) :
m_coord( rhs.m_coord ),
m_value( rhs.m_value ),
m_dim( rhs.m_dim ),
m_flops( rhs.m_flops ) {}
inline
CooProductTensor & operator = ( const CooProductTensor & rhs )
{
m_coord = rhs.m_coord;
m_value = rhs.m_value;
m_dim = rhs.m_dim;
m_flops = rhs.m_flops;
return *this;
}
/** \brief Dimension of the tensor. */
KOKKOS_INLINE_FUNCTION
size_type dimension() const { return m_dim; }
/** \brief Number of sparse entries. */
KOKKOS_INLINE_FUNCTION
size_type entry_count() const { return m_coord.dimension_0(); }
/** \brief Get (i,j,k) coordinates of an entry */
KOKKOS_INLINE_FUNCTION
void coord( const size_type entry,
size_type& i, size_type& j, size_type& k ) const {
i = m_coord(entry,0);
j = m_coord(entry,1);
k = m_coord(entry,2);
}
/** \brief Value of an entry */
KOKKOS_INLINE_FUNCTION
const value_type & value( const size_type entry ) const
{ return m_value( entry ); }
/** \brief Number of non-zero's */
KOKKOS_INLINE_FUNCTION
size_type num_non_zeros() const { return m_coord.dimension_0(); }
/** \brief Number flop's per multiply-add */
KOKKOS_INLINE_FUNCTION
size_type num_flops() const { return m_flops; }
template <typename OrdinalType>
static CooProductTensor
create( const Stokhos::ProductBasis<OrdinalType,ValueType>& basis,
const Stokhos::Sparse3Tensor<OrdinalType,ValueType>& Cijk,
const Teuchos::ParameterList& params = Teuchos::ParameterList())
{
typedef Stokhos::Sparse3Tensor<OrdinalType,ValueType> Cijk_type;
typedef typename Cijk_type::i_iterator i_iterator;
typedef typename Cijk_type::ik_iterator k_iterator;
typedef typename Cijk_type::ikj_iterator j_iterator;
// Compute entry count
size_type entry_count = 0;
for (i_iterator i_it = Cijk.i_begin(); i_it!=Cijk.i_end(); ++i_it) {
for (k_iterator k_it = Cijk.k_begin(i_it); k_it != Cijk.k_end(i_it);
++k_it) {
OrdinalType k = index(k_it);
for (j_iterator j_it = Cijk.j_begin(k_it); j_it != Cijk.j_end(k_it);
++j_it) {
OrdinalType j = index(j_it);
if (j >= k) {
++entry_count;
}
}
}
}
// Align entry_count
#if defined( KOKKOS_HAVE_CUDA )
enum { Align = Kokkos::Impl::is_same<ExecutionSpace,Kokkos::Cuda>::value ? 32 : 1 };
#else
enum { Align = 1 };
#endif
entry_count = (entry_count+Align-1) & ~(Align-1);
TEUCHOS_ASSERT(entry_count % Align == 0);
// Allocate tensor data
CooProductTensor tensor;
tensor.m_coord = coord_array_type( "tensor_coord", entry_count );
tensor.m_value = value_array_type( "tensor_value", entry_count );
tensor.m_dim = basis.size();
tensor.m_flops = 5*entry_count + tensor.m_dim;
// Create mirror, is a view if is host memory
typename coord_array_type::HostMirror
host_coord = Kokkos::create_mirror_view( tensor.m_coord );
typename value_array_type::HostMirror
host_value = Kokkos::create_mirror_view( tensor.m_value );
// Set entries
size_type n = 0;
OrdinalType i=0, j=0, k=0;
for (i_iterator i_it = Cijk.i_begin(); i_it!=Cijk.i_end(); ++i_it) {
i = index(i_it);
for (k_iterator k_it = Cijk.k_begin(i_it); k_it != Cijk.k_end(i_it);
++k_it) {
k = index(k_it);
for (j_iterator j_it = Cijk.j_begin(k_it); j_it != Cijk.j_end(k_it);
++j_it) {
j = index(j_it);
ValueType c = Stokhos::value(j_it);
if (j >= k) {
host_value(n) = (j != k) ? c : 0.5*c;
host_coord(n,0) = i;
host_coord(n,1) = j;
host_coord(n,2) = k;
++n;
}
}
}
}
for (; n < entry_count; ++n) {
host_value(n) = 0.0;
host_coord(n,0) = i;
host_coord(n,1) = j;
host_coord(n,2) = k;
}
// Copy data to device if necessary
Kokkos::deep_copy( tensor.m_coord, host_coord );
Kokkos::deep_copy( tensor.m_value, host_value );
return tensor;
}
void print(std::ostream& os) const {
size_type num_entry = entry_count();
typename coord_array_type::HostMirror
host_coord = Kokkos::create_mirror_view( m_coord );
typename value_array_type::HostMirror
host_value = Kokkos::create_mirror_view( m_value );
Kokkos::deep_copy( host_coord, m_coord );
Kokkos::deep_copy( host_value, m_value );
os << "CooProductTensor: dim = "
<< dimension() << ", entry_count = "
<< num_entry << std::endl
<< "Entries: i j k : cijk" << std::endl;
for (size_type l=0; l<num_entry; ++l) {
size_type i = host_coord(l,0);
size_type j = host_coord(l,1);
size_type k = host_coord(l,2);
ValueType cijk = host_value(l);
os << "\t " << l << " : " << i << " " << j << " " << k << " = " << cijk
<< std::endl;
if (l > 0 && l % 32 == 0) os << std::endl;
}
}
};
template< class Device, bool Pack, typename OrdinalType, typename ValueType >
CooProductTensor<ValueType, Device, Pack>
create_coo_product_tensor(
const Stokhos::ProductBasis<OrdinalType,ValueType>& basis,
const Stokhos::Sparse3Tensor<OrdinalType,ValueType>& Cijk,
const Teuchos::ParameterList& params = Teuchos::ParameterList())
{
return CooProductTensor<ValueType, Device, Pack>::create(
basis, Cijk, params );
}
template <typename ValueType, typename Device, bool Pack>
std::ostream& operator << (
std::ostream& os, const CooProductTensor<ValueType,Device,Pack>& tensor)
{
tensor.print(os);
return os;
}
template< typename ValueType, typename Device, bool Pack >
class BlockMultiply< CooProductTensor< ValueType, Device, Pack > >
{
public:
typedef typename Device::size_type size_type;
typedef CooProductTensor< ValueType, Device, Pack > tensor_type;
template< typename MatrixValue , typename VectorValue >
KOKKOS_INLINE_FUNCTION
static void apply( const tensor_type & tensor,
const MatrixValue * const a,
const VectorValue * const x,
VectorValue * const y )
{
const size_type nEntry = tensor.entry_count();
size_type i = 0, j = 0, k = 0, i_prev = -1;
VectorValue val = 0.0, carry_val = 0.0;
for ( size_type entry = 0 ; entry < nEntry ; ++entry ) {
tensor.coord(entry, i, j, k);
val = tensor.value(entry) * ( a[j] * x[k] + a[k] * x[j] );
if (i == i_prev)
carry_val += val;
else {
y[i_prev] += carry_val;
carry_val = val;
}
i_prev = i;
}
y[i] += carry_val;
}
KOKKOS_INLINE_FUNCTION
static size_type matrix_size( const tensor_type & tensor )
{ return tensor.dimension(); }
KOKKOS_INLINE_FUNCTION
static size_type vector_size( const tensor_type & tensor )
{ return tensor.dimension(); }
};
} /* namespace Stokhos */
//----------------------------------------------------------------------------
//----------------------------------------------------------------------------
#endif /* #ifndef STOKHOS_COO_PRODUCT_TENSOR_HPP */
|