Blame view
tools/cub-1.8.0/test/test_block_radix_sort.cu
24.8 KB
8dcb6dfcb
first commit
|
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 |
/******************************************************************************
* Copyright (c) 2011, Duane Merrill. All rights reserved.
* Copyright (c) 2011-2018, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * 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.
* * Neither the name of the NVIDIA CORPORATION 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 NVIDIA CORPORATION 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.
*
******************************************************************************/
/******************************************************************************
* Test of BlockRadixSort utilities
******************************************************************************/
// Ensure printing of CUDA runtime errors to console
#define CUB_STDERR
#include <stdio.h>
#include <algorithm>
#include <iostream>
#include <cub/block/block_radix_sort.cuh>
#include <cub/block/block_load.cuh>
#include <cub/block/block_store.cuh>
#include <cub/util_allocator.cuh>
#include "test_util.h"
using namespace cub;
//---------------------------------------------------------------------
// Globals, constants and typedefs
//---------------------------------------------------------------------
bool g_verbose = false;
CachingDeviceAllocator g_allocator(true);
//---------------------------------------------------------------------
// Test kernels
//---------------------------------------------------------------------
/// Specialized descending, blocked -> blocked
template <int BLOCK_THREADS, typename BlockRadixSort, int ITEMS_PER_THREAD, typename Key, typename Value>
__device__ __forceinline__ void TestBlockSort(
typename BlockRadixSort::TempStorage &temp_storage,
Key (&keys)[ITEMS_PER_THREAD],
Value (&values)[ITEMS_PER_THREAD],
Key *d_keys,
Value *d_values,
int begin_bit,
int end_bit,
clock_t &stop,
Int2Type<true> is_descending,
Int2Type<true> is_blocked_output)
{
BlockRadixSort(temp_storage).SortDescending(keys, values, begin_bit, end_bit);
stop = clock();
StoreDirectBlocked(threadIdx.x, d_keys, keys);
StoreDirectBlocked(threadIdx.x, d_values, values);
}
/// Specialized descending, blocked -> striped
template <int BLOCK_THREADS, typename BlockRadixSort, int ITEMS_PER_THREAD, typename Key, typename Value>
__device__ __forceinline__ void TestBlockSort(
typename BlockRadixSort::TempStorage &temp_storage,
Key (&keys)[ITEMS_PER_THREAD],
Value (&values)[ITEMS_PER_THREAD],
Key *d_keys,
Value *d_values,
int begin_bit,
int end_bit,
clock_t &stop,
Int2Type<true> is_descending,
Int2Type<false> is_blocked_output)
{
BlockRadixSort(temp_storage).SortDescendingBlockedToStriped(keys, values, begin_bit, end_bit);
stop = clock();
StoreDirectStriped<BLOCK_THREADS>(threadIdx.x, d_keys, keys);
StoreDirectStriped<BLOCK_THREADS>(threadIdx.x, d_values, values);
}
/// Specialized ascending, blocked -> blocked
template <int BLOCK_THREADS, typename BlockRadixSort, int ITEMS_PER_THREAD, typename Key, typename Value>
__device__ __forceinline__ void TestBlockSort(
typename BlockRadixSort::TempStorage &temp_storage,
Key (&keys)[ITEMS_PER_THREAD],
Value (&values)[ITEMS_PER_THREAD],
Key *d_keys,
Value *d_values,
int begin_bit,
int end_bit,
clock_t &stop,
Int2Type<false> is_descending,
Int2Type<true> is_blocked_output)
{
BlockRadixSort(temp_storage).Sort(keys, values, begin_bit, end_bit);
stop = clock();
StoreDirectBlocked(threadIdx.x, d_keys, keys);
StoreDirectBlocked(threadIdx.x, d_values, values);
}
/// Specialized ascending, blocked -> striped
template <int BLOCK_THREADS, typename BlockRadixSort, int ITEMS_PER_THREAD, typename Key, typename Value>
__device__ __forceinline__ void TestBlockSort(
typename BlockRadixSort::TempStorage &temp_storage,
Key (&keys)[ITEMS_PER_THREAD],
Value (&values)[ITEMS_PER_THREAD],
Key *d_keys,
Value *d_values,
int begin_bit,
int end_bit,
clock_t &stop,
Int2Type<false> is_descending,
Int2Type<false> is_blocked_output)
{
BlockRadixSort(temp_storage).SortBlockedToStriped(keys, values, begin_bit, end_bit);
stop = clock();
StoreDirectStriped<BLOCK_THREADS>(threadIdx.x, d_keys, keys);
StoreDirectStriped<BLOCK_THREADS>(threadIdx.x, d_values, values);
}
/**
* BlockRadixSort kernel
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS,
bool MEMOIZE_OUTER_SCAN,
BlockScanAlgorithm INNER_SCAN_ALGORITHM,
cudaSharedMemConfig SMEM_CONFIG,
int DESCENDING,
int BLOCKED_OUTPUT,
typename Key,
typename Value>
__launch_bounds__ (BLOCK_THREADS, 1)
__global__ void Kernel(
Key *d_keys,
Value *d_values,
int begin_bit,
int end_bit,
clock_t *d_elapsed)
{
// Threadblock load/store abstraction types
typedef BlockRadixSort<
Key,
BLOCK_THREADS,
ITEMS_PER_THREAD,
Value,
RADIX_BITS,
MEMOIZE_OUTER_SCAN,
INNER_SCAN_ALGORITHM,
SMEM_CONFIG>
BlockRadixSortT;
// Allocate temp storage in shared memory
__shared__ typename BlockRadixSortT::TempStorage temp_storage;
// Items per thread
Key keys[ITEMS_PER_THREAD];
Value values[ITEMS_PER_THREAD];
LoadDirectBlocked(threadIdx.x, d_keys, keys);
LoadDirectBlocked(threadIdx.x, d_values, values);
// Start cycle timer
clock_t stop;
clock_t start = clock();
TestBlockSort<BLOCK_THREADS, BlockRadixSortT>(
temp_storage, keys, values, d_keys, d_values, begin_bit, end_bit, stop, Int2Type<DESCENDING>(), Int2Type<BLOCKED_OUTPUT>());
// Store time
if (threadIdx.x == 0)
*d_elapsed = (start > stop) ? start - stop : stop - start;
}
//---------------------------------------------------------------------
// Host testing subroutines
//---------------------------------------------------------------------
/**
* Simple key-value pairing
*/
template <
typename Key,
typename Value,
bool IS_FLOAT = (Traits<Key>::CATEGORY == FLOATING_POINT)>
struct Pair
{
Key key;
Value value;
bool operator<(const Pair &b) const
{
return (key < b.key);
}
};
/**
* Simple key-value pairing (specialized for floating point types)
*/
template <typename Key, typename Value>
struct Pair<Key, Value, true>
{
Key key;
Value value;
bool operator<(const Pair &b) const
{
if (key < b.key)
return true;
if (key > b.key)
return false;
// Key in unsigned bits
typedef typename Traits<Key>::UnsignedBits UnsignedBits;
// Return true if key is negative zero and b.key is positive zero
UnsignedBits key_bits = *reinterpret_cast<UnsignedBits*>(const_cast<Key*>(&key));
UnsignedBits b_key_bits = *reinterpret_cast<UnsignedBits*>(const_cast<Key*>(&b.key));
UnsignedBits HIGH_BIT = Traits<Key>::HIGH_BIT;
return ((key_bits & HIGH_BIT) != 0) && ((b_key_bits & HIGH_BIT) == 0);
}
};
/**
* Initialize key-value sorting problem.
*/
template <bool DESCENDING, typename Key, typename Value>
void Initialize(
GenMode gen_mode,
Key *h_keys,
Value *h_values,
Key *h_reference_keys,
Value *h_reference_values,
int num_items,
int entropy_reduction,
int begin_bit,
int end_bit)
{
Pair<Key, Value> *h_pairs = new Pair<Key, Value>[num_items];
for (int i = 0; i < num_items; ++i)
{
InitValue(gen_mode, h_keys[i], i);
RandomBits(h_values[i]);
// Mask off unwanted portions
int num_bits = end_bit - begin_bit;
if ((begin_bit > 0) || (end_bit < sizeof(Key) * 8))
{
unsigned long long base = 0;
memcpy(&base, &h_keys[i], sizeof(Key));
base &= ((1ull << num_bits) - 1) << begin_bit;
memcpy(&h_keys[i], &base, sizeof(Key));
}
h_pairs[i].key = h_keys[i];
h_pairs[i].value = h_values[i];
}
if (DESCENDING) std::reverse(h_pairs, h_pairs + num_items);
std::stable_sort(h_pairs, h_pairs + num_items);
if (DESCENDING) std::reverse(h_pairs, h_pairs + num_items);
for (int i = 0; i < num_items; ++i)
{
h_reference_keys[i] = h_pairs[i].key;
h_reference_values[i] = h_pairs[i].value;
}
delete[] h_pairs;
}
/**
* Test BlockRadixSort kernel
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS,
bool MEMOIZE_OUTER_SCAN,
BlockScanAlgorithm INNER_SCAN_ALGORITHM,
cudaSharedMemConfig SMEM_CONFIG,
bool DESCENDING,
bool BLOCKED_OUTPUT,
typename Key,
typename Value>
void TestDriver(
GenMode gen_mode,
int entropy_reduction,
int begin_bit,
int end_bit)
{
enum
{
TILE_SIZE = BLOCK_THREADS * ITEMS_PER_THREAD,
KEYS_ONLY = Equals<Value, NullType>::VALUE,
};
// Allocate host arrays
Key *h_keys = new Key[TILE_SIZE];
Key *h_reference_keys = new Key[TILE_SIZE];
Value *h_values = new Value[TILE_SIZE];
Value *h_reference_values = new Value[TILE_SIZE];
// Allocate device arrays
Key *d_keys = NULL;
Value *d_values = NULL;
clock_t *d_elapsed = NULL;
CubDebugExit(g_allocator.DeviceAllocate((void**)&d_keys, sizeof(Key) * TILE_SIZE));
CubDebugExit(g_allocator.DeviceAllocate((void**)&d_values, sizeof(Value) * TILE_SIZE));
CubDebugExit(g_allocator.DeviceAllocate((void**)&d_elapsed, sizeof(clock_t)));
// Initialize problem and solution on host
Initialize<DESCENDING>(gen_mode, h_keys, h_values, h_reference_keys, h_reference_values,
TILE_SIZE, entropy_reduction, begin_bit, end_bit);
// Copy problem to device
CubDebugExit(cudaMemcpy(d_keys, h_keys, sizeof(Key) * TILE_SIZE, cudaMemcpyHostToDevice));
CubDebugExit(cudaMemcpy(d_values, h_values, sizeof(Value) * TILE_SIZE, cudaMemcpyHostToDevice));
printf("%s "
"BLOCK_THREADS(%d) "
"ITEMS_PER_THREAD(%d) "
"RADIX_BITS(%d) "
"MEMOIZE_OUTER_SCAN(%d) "
"INNER_SCAN_ALGORITHM(%d) "
"SMEM_CONFIG(%d) "
"DESCENDING(%d) "
"BLOCKED_OUTPUT(%d) "
"sizeof(Key)(%d) "
"sizeof(Value)(%d) "
"gen_mode(%d), "
"entropy_reduction(%d) "
"begin_bit(%d) "
"end_bit(%d), "
"samples(%d)
",
((KEYS_ONLY) ? "Keys-only" : "Key-value"),
BLOCK_THREADS,
ITEMS_PER_THREAD,
RADIX_BITS,
MEMOIZE_OUTER_SCAN,
INNER_SCAN_ALGORITHM,
SMEM_CONFIG,
DESCENDING,
BLOCKED_OUTPUT,
(int) sizeof(Key),
(int) sizeof(Value),
gen_mode,
entropy_reduction,
begin_bit,
end_bit,
g_num_rand_samples);
// Set shared memory config
cudaDeviceSetSharedMemConfig(SMEM_CONFIG);
// Run kernel
Kernel<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, SMEM_CONFIG, DESCENDING, BLOCKED_OUTPUT><<<1, BLOCK_THREADS>>>(
d_keys, d_values, begin_bit, end_bit, d_elapsed);
// Flush kernel output / errors
CubDebugExit(cudaPeekAtLastError());
CubDebugExit(cudaDeviceSynchronize());
// Check keys results
printf("\tKeys: ");
int compare = CompareDeviceResults(h_reference_keys, d_keys, TILE_SIZE, g_verbose, g_verbose);
printf("%s
", compare ? "FAIL" : "PASS");
AssertEquals(0, compare);
// Check value results
if (!KEYS_ONLY)
{
printf("\tValues: ");
int compare = CompareDeviceResults(h_reference_values, d_values, TILE_SIZE, g_verbose, g_verbose);
printf("%s
", compare ? "FAIL" : "PASS");
AssertEquals(0, compare);
}
printf("
");
printf("\tElapsed clocks: ");
DisplayDeviceResults(d_elapsed, 1);
printf("
");
// Cleanup
if (h_keys) delete[] h_keys;
if (h_reference_keys) delete[] h_reference_keys;
if (h_values) delete[] h_values;
if (h_reference_values) delete[] h_reference_values;
if (d_keys) CubDebugExit(g_allocator.DeviceFree(d_keys));
if (d_values) CubDebugExit(g_allocator.DeviceFree(d_values));
if (d_elapsed) CubDebugExit(g_allocator.DeviceFree(d_elapsed));
}
/**
* Test driver (valid tile size <= MAX_SMEM_BYTES)
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS,
bool MEMOIZE_OUTER_SCAN,
BlockScanAlgorithm INNER_SCAN_ALGORITHM,
cudaSharedMemConfig SMEM_CONFIG,
bool DESCENDING,
bool BLOCKED_OUTPUT,
typename Key,
typename Value>
void TestValid(Int2Type<true> fits_smem_capacity)
{
// Iterate begin_bit
for (int begin_bit = 0; begin_bit <= 1; begin_bit++)
{
// Iterate end bit
for (int end_bit = begin_bit + 1; end_bit <= sizeof(Key) * 8; end_bit = end_bit * 2 + begin_bit)
{
// Uniform key distribution
TestDriver<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, SMEM_CONFIG, DESCENDING, BLOCKED_OUTPUT, Key, Value>(
UNIFORM, 0, begin_bit, end_bit);
// Sequential key distribution
TestDriver<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, SMEM_CONFIG, DESCENDING, BLOCKED_OUTPUT, Key, Value>(
INTEGER_SEED, 0, begin_bit, end_bit);
// Iterate random with entropy_reduction
for (int entropy_reduction = 0; entropy_reduction <= 9; entropy_reduction += 3)
{
TestDriver<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, SMEM_CONFIG, DESCENDING, BLOCKED_OUTPUT, Key, Value>(
RANDOM, entropy_reduction, begin_bit, end_bit);
}
}
}
}
/**
* Test driver (invalid tile size)
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS,
bool MEMOIZE_OUTER_SCAN,
BlockScanAlgorithm INNER_SCAN_ALGORITHM,
cudaSharedMemConfig SMEM_CONFIG,
bool DESCENDING,
bool BLOCKED_OUTPUT,
typename Key,
typename Value>
void TestValid(Int2Type<false> fits_smem_capacity)
{}
/**
* Test ascending/descending and to-blocked/to-striped
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS,
bool MEMOIZE_OUTER_SCAN,
BlockScanAlgorithm INNER_SCAN_ALGORITHM,
cudaSharedMemConfig SMEM_CONFIG,
typename Key,
typename Value>
void Test()
{
// Check size of smem storage for the target arch to make sure it will fit
typedef BlockRadixSort<Key, BLOCK_THREADS, ITEMS_PER_THREAD, Value, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, SMEM_CONFIG> BlockRadixSortT;
#if defined(SM100) || defined(SM110) || defined(SM130)
Int2Type<sizeof(typename BlockRadixSortT::TempStorage) <= 16 * 1024> fits_smem_capacity;
#else
Int2Type<(sizeof(typename BlockRadixSortT::TempStorage) <= 48 * 1024)> fits_smem_capacity;
#endif
// Sort-ascending, to-striped
TestValid<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, SMEM_CONFIG, true, false, Key, Value>(fits_smem_capacity);
// Sort-descending, to-blocked
TestValid<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, SMEM_CONFIG, false, true, Key, Value>(fits_smem_capacity);
// Not necessary
// TestValid<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, SMEM_CONFIG, false, false, Key, Value>(fits_smem_capacity);
// TestValid<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, SMEM_CONFIG, true, true, Key, Value>(fits_smem_capacity);
}
/**
* Test value type and smem config
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS,
bool MEMOIZE_OUTER_SCAN,
BlockScanAlgorithm INNER_SCAN_ALGORITHM,
typename Key>
void TestKeys()
{
// Test keys-only sorting with both smem configs
Test<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, cudaSharedMemBankSizeFourByte, Key, NullType>(); // Keys-only (4-byte smem bank config)
#if !defined(SM100) && !defined(SM110) && !defined(SM130) && !defined(SM200)
Test<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, cudaSharedMemBankSizeEightByte, Key, NullType>(); // Keys-only (8-byte smem bank config)
#endif
}
/**
* Test value type and smem config
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS,
bool MEMOIZE_OUTER_SCAN,
BlockScanAlgorithm INNER_SCAN_ALGORITHM,
typename Key>
void TestKeysAndPairs()
{
// Test pairs sorting with only 4-byte configs
Test<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, cudaSharedMemBankSizeFourByte, Key, char>(); // With small-values
Test<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, cudaSharedMemBankSizeFourByte, Key, Key>(); // With same-values
Test<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, cudaSharedMemBankSizeFourByte, Key, TestFoo>(); // With large values
}
/**
* Test key type
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS,
bool MEMOIZE_OUTER_SCAN,
BlockScanAlgorithm INNER_SCAN_ALGORITHM>
void Test()
{
// Get ptx version
int ptx_version;
CubDebugExit(PtxVersion(ptx_version));
#ifdef TEST_KEYS_ONLY
// Test unsigned types with keys-only
TestKeys<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, unsigned char>();
TestKeys<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, unsigned short>();
TestKeys<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, unsigned int>();
TestKeys<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, unsigned long>();
TestKeys<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, unsigned long long>();
#else
// Test signed and fp types with paired values
TestKeysAndPairs<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, char>();
TestKeysAndPairs<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, short>();
TestKeysAndPairs<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, int>();
TestKeysAndPairs<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, long>();
TestKeysAndPairs<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, long long>();
TestKeysAndPairs<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, float>();
if (ptx_version > 120)
{
// Don't check doubles on PTX120 or below because they're down-converted
TestKeysAndPairs<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, INNER_SCAN_ALGORITHM, double>();
}
#endif
}
/**
* Test inner scan algorithm
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS,
bool MEMOIZE_OUTER_SCAN>
void Test()
{
Test<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, BLOCK_SCAN_RAKING>();
Test<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, MEMOIZE_OUTER_SCAN, BLOCK_SCAN_WARP_SCANS>();
}
/**
* Test outer scan algorithm
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD,
int RADIX_BITS>
void Test()
{
Test<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, true>();
Test<BLOCK_THREADS, ITEMS_PER_THREAD, RADIX_BITS, false>();
}
/**
* Test radix bits
*/
template <
int BLOCK_THREADS,
int ITEMS_PER_THREAD>
void Test()
{
Test<BLOCK_THREADS, ITEMS_PER_THREAD, 1>();
Test<BLOCK_THREADS, ITEMS_PER_THREAD, 2>();
Test<BLOCK_THREADS, ITEMS_PER_THREAD, 5>();
}
/**
* Test items per thread
*/
template <int BLOCK_THREADS>
void Test()
{
Test<BLOCK_THREADS, 1>();
#if defined(SM100) || defined(SM110) || defined(SM130)
// Open64 compiler can't handle the number of test cases
#else
Test<BLOCK_THREADS, 4>();
#endif
Test<BLOCK_THREADS, 11>();
}
/**
* Main
*/
int main(int argc, char** argv)
{
// Initialize command line
CommandLineArgs args(argc, argv);
g_verbose = args.CheckCmdLineFlag("v");
// Print usage
if (args.CheckCmdLineFlag("help"))
{
printf("%s "
"[--device=<device-id>] "
"[--v] "
"
", argv[0]);
exit(0);
}
// Initialize device
CubDebugExit(args.DeviceInit());
#ifdef QUICK_TEST
{
typedef float T;
TestDriver<32, 4, 4, true, BLOCK_SCAN_WARP_SCANS, cudaSharedMemBankSizeFourByte, false, false, T, NullType>(INTEGER_SEED, 0, 0, sizeof(T) * 8);
}
/*
// Compile/run quick tests
typedef unsigned int T;
TestDriver<64, 17, 4, true, BLOCK_SCAN_WARP_SCANS, cudaSharedMemBankSizeFourByte, false, false, T, NullType>(RANDOM, 0, 0, sizeof(T) * 8);
TestDriver<96, 8, 4, true, BLOCK_SCAN_WARP_SCANS, cudaSharedMemBankSizeFourByte, false, false, T, NullType>(RANDOM, 0, 0, sizeof(T) * 8);
TestDriver<128, 2, 4, true, BLOCK_SCAN_WARP_SCANS, cudaSharedMemBankSizeFourByte, false, false, T, NullType>(RANDOM, 0, 0, sizeof(T) * 8);
*/
#else
// Compile/run thorough tests
Test<32>();
Test<64>();
Test<160>();
#endif // QUICK_TEST
return 0;
}
|