Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[GPU-Plugin] Improved load balancing search (#2521)

Browse Source
This commit is contained in:
Rory Mitchell 2017-07-17 11:50:57 +12:00 committed by GitHub
parent 33ee7d1615
commit c85bf9859e
2 changed files with 170 additions and 109 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

197
plugin/updater_gpu/src/device_helpers.cuh
View File

@ -132,9 +132,12 @@ struct Timer {
void reset() { start = ClockT::now(); } void reset() { start = ClockT::now(); }
int64_t elapsed() const { return (ClockT::now() - start).count(); } int64_t elapsed() const { return (ClockT::now() - start).count(); }
double elapsedSeconds() const {
return elapsed() * ((double)ClockT::period::num / ClockT::period::den);
}
void printElapsed(std::string label) { void printElapsed(std::string label) {
// synchronize_n_devices(n_devices, dList); // synchronize_n_devices(n_devices, dList);
printf("%s:\t %lld\n", label.c_str(), elapsed()); printf("%s:\t %fs\n", label.c_str(), elapsedSeconds());
reset(); reset();
} }
}; };
@ -650,116 +653,124 @@ struct BernoulliRng {
// Load balancing search // Load balancing search
template <typename func_t> template <typename coordinate_t, typename segments_t, typename offset_t>
class LauncherItr { void FindMergePartitions(int device_idx, coordinate_t *d_tile_coordinates, int num_tiles,
public: int tile_size, segments_t segments, offset_t num_rows,
int idx; offset_t num_elements) {
func_t f; dh::launch_n(device_idx, num_tiles + 1, [=] __device__(int idx) {
XGBOOST_DEVICE LauncherItr() : idx(0) {} offset_t diagonal = idx * tile_size;
XGBOOST_DEVICE LauncherItr(int idx, func_t f) : idx(idx), f(f) {} coordinate_t tile_coordinate;
XGBOOST_DEVICE LauncherItr &operator=(int output) { cub::CountingInputIterator<offset_t> nonzero_indices(0);
f(idx, output);
return *this;
}
};
template <typename func_t> // Search the merge path
// Cast to signed integer as this function can have negatives
cub::MergePathSearch(static_cast<int64_t>(diagonal), segments + 1,
nonzero_indices, static_cast<int64_t>(num_rows),
static_cast<int64_t>(num_elements), tile_coordinate);
// Output starting offset
d_tile_coordinates[idx] = tile_coordinate;
});
}
template <int TILE_SIZE, int ITEMS_PER_THREAD, int BLOCK_THREADS,
typename offset_t, typename coordinate_t, typename func_t,
typename segments_iter>
__global__ void LbsKernel(coordinate_t *d_coordinates,
segments_iter segment_end_offsets, func_t f,
offset_t num_segments) {
int tile = blockIdx.x;
coordinate_t tile_start_coord = d_coordinates[tile];
coordinate_t tile_end_coord = d_coordinates[tile + 1];
int64_t tile_num_rows = tile_end_coord.x - tile_start_coord.x;
int64_t tile_num_elements = tile_end_coord.y - tile_start_coord.y;
cub::CountingInputIterator<offset_t> tile_element_indices(tile_start_coord.y);
coordinate_t thread_start_coord;
typedef typename std::iterator_traits<segments_iter>::value_type segment_t;
__shared__ struct {
segment_t tile_segment_end_offsets[TILE_SIZE + 1];
segment_t output_segment[TILE_SIZE];
} temp_storage;
for (auto item : dh::block_stride_range(int(0), int(tile_num_rows + 1))) {
temp_storage.tile_segment_end_offsets[item] =
segment_end_offsets[min(tile_start_coord.x + item, num_segments - 1)];
}
__syncthreads();
int64_t diag = threadIdx.x * ITEMS_PER_THREAD;
// Cast to signed integer as this function can have negatives
cub::MergePathSearch(diag, // Diagonal
temp_storage.tile_segment_end_offsets, // List A
tile_element_indices, // List B
tile_num_rows, tile_num_elements, thread_start_coord);
coordinate_t thread_current_coord = thread_start_coord;
#pragma unroll
for (int ITEM = 0; ITEM < ITEMS_PER_THREAD; ++ITEM) {
if (tile_element_indices[thread_current_coord.y] <
temp_storage.tile_segment_end_offsets[thread_current_coord.x]) {
temp_storage.output_segment[thread_current_coord.y] =
thread_current_coord.x + tile_start_coord.x;
++thread_current_coord.y;
} else {
++thread_current_coord.x;
}
}
__syncthreads();
for (auto item : dh::block_stride_range(int(0), int(tile_num_elements))) {
f(tile_start_coord.y + item, temp_storage.output_segment[item]);
}
}
/** /**
* \class DiscardLambdaItr * \fn template <typename func_t, typename segments_iter, typename offset_t>
* void TransformLbs(int device_idx, dh::CubMemory *temp_memory, offset_t count,
* segments_iter segments, offset_t num_segments, func_t f)
* *
* \brief Thrust compatible iterator type - discards algorithm output and * \brief Load balancing search function. Reads a CSR type matrix description
* launches device lambda with the index of the output and the algorithm output as arguments. * and allows a function to be executed on each element. Search 'modern GPU load
* * balancing search' for more information.
* \author Rory
* \date 7/9/2017
*/
class DiscardLambdaItr {
public:
// Required iterator traits
typedef DiscardLambdaItr self_type; ///< My own type
typedef ptrdiff_t
difference_type; ///< Type to express the result of subtracting
/// one iterator from another
typedef LauncherItr<func_t>
value_type; ///< The type of the element the iterator can point to
typedef value_type *pointer; ///< The type of a pointer to an element the
/// iterator can point to
typedef value_type reference; ///< The type of a reference to an element the
/// iterator can point to
typedef typename thrust::detail::iterator_facade_category<
thrust::any_system_tag, thrust::random_access_traversal_tag, value_type,
reference>::type iterator_category; ///< The iterator category
private:
difference_type offset;
func_t f;
public:
XGBOOST_DEVICE DiscardLambdaItr(func_t f) : offset(0), f(f) {}
XGBOOST_DEVICE DiscardLambdaItr(difference_type offset, func_t f)
: offset(offset), f(f) {}
XGBOOST_DEVICE self_type operator+(const int &b) const {
return DiscardLambdaItr(offset + b, f);
}
XGBOOST_DEVICE self_type operator++() {
offset++;
return *this;
}
XGBOOST_DEVICE self_type operator++(int) {
self_type retval = *this;
offset++;
return retval;
}
XGBOOST_DEVICE self_type &operator+=(const int &b) {
offset += b;
return *this;
}
XGBOOST_DEVICE reference operator*() const {
return LauncherItr<func_t>(offset, f);
}
XGBOOST_DEVICE reference operator[](int idx) {
self_type offset = (*this) + idx;
return *offset;
}
};
/**
* \fn template <typename func_t, typename segments_t> void TransformLbs(int device_idx, dh::CubMemory *temp_memory, int count, thrust::device_ptr<segments_t> segments, int num_segments, func_t f)
*
* \brief Load balancing search function. Reads a CSR type matrix description and allows a function
* to be executed on each element. Search 'modern GPU load balancing search for more
* information'.
* *
* \author Rory * \author Rory
* \date 7/9/2017 * \date 7/9/2017
* *
* \tparam segments_t Type of the segments t. * \tparam func_t Type of the function t.
* \tparam segments_iter Type of the segments iterator.
* \tparam offset_t Type of the offset.
* \tparam segments_t Type of the segments t.
* \param device_idx Zero-based index of the device. * \param device_idx Zero-based index of the device.
* \param [in,out] temp_memory Temporary memory allocator. * \param [in,out] temp_memory Temporary memory allocator.
* \param count Number of elements. * \param count Number of elements.
* \param segments Device pointed to segments. * \param segments Device pointer to segments.
* \param num_segments Number of segments. * \param num_segments Number of segments.
* \param f Lambda to be executed on matrix elements. * \param f Lambda to be executed on matrix elements.
*/ */
template <typename func_t, typename segments_t> template <typename func_t, typename segments_iter, typename offset_t>
void TransformLbs(int device_idx, dh::CubMemory *temp_memory, int count, void TransformLbs(int device_idx, dh::CubMemory *temp_memory, offset_t count,
thrust::device_ptr<segments_t> segments, int num_segments, segments_iter segments, offset_t num_segments, func_t f) {
func_t f) { typedef typename cub::CubVector<offset_t, 2>::Type coordinate_t;
safe_cuda(cudaSetDevice(device_idx)); dh::safe_cuda(cudaSetDevice(device_idx));
auto counting = thrust::make_counting_iterator(0); const int BLOCK_THREADS = 256;
const int ITEMS_PER_THREAD = 1;
const int TILE_SIZE = BLOCK_THREADS * ITEMS_PER_THREAD;
int num_tiles = dh::div_round_up(count + num_segments, BLOCK_THREADS);
auto f_wrapper = [=] __device__(int idx, int upper_bound) { temp_memory->LazyAllocate(sizeof(coordinate_t) * (num_tiles + 1));
f(idx, upper_bound - 1); coordinate_t *tmp_tile_coordinates =
}; reinterpret_cast<coordinate_t *>(temp_memory->d_temp_storage);
DiscardLambdaItr<decltype(f_wrapper)> itr(f_wrapper); FindMergePartitions(device_idx, tmp_tile_coordinates, num_tiles, BLOCK_THREADS, segments,
num_segments, count);
thrust::upper_bound(thrust::cuda::par(*temp_memory), segments, LbsKernel<TILE_SIZE, ITEMS_PER_THREAD, BLOCK_THREADS, offset_t>
segments + num_segments, counting, counting + count, itr); <<<num_tiles, BLOCK_THREADS>>>(tmp_tile_coordinates, segments + 1, f,
num_segments);
} }
} // namespace dh } // namespace dh

82
plugin/updater_gpu/test/cpp/test_device_helpers.cu
View File

@ -7,22 +7,72 @@
#include "../../src/device_helpers.cuh" #include "../../src/device_helpers.cuh"
#include "gtest/gtest.h" #include "gtest/gtest.h"
static const std::vector<int> gidx = {0, 2, 5, 1, 3, 6, 0, 2, 0, 7}; void CreateTestData(xgboost::bst_uint num_rows, int max_row_size,
static const std::vector<int> row_ptr = {0, 3, 6, 8, 10}; thrust::host_vector<int> *row_ptr,
static const std::vector<int> lbs_seg_output = {0, 0, 0, 1, 1, 1, 2, 2, 3, 3}; thrust::host_vector<xgboost::bst_uint> *rows) {
row_ptr->resize(num_rows + 1);
int sum = 0;
for (int i = 0; i <= num_rows; i++) {
(*row_ptr)[i] = sum;
sum += rand() % max_row_size; // NOLINT
thrust::device_vector<int> test_lbs() { if (i < num_rows) {
thrust::device_vector<int> device_gidx = gidx; for (int j = (*row_ptr)[i]; j < sum; j++) {
thrust::device_vector<int> device_row_ptr = row_ptr; (*rows).push_back(i);
thrust::device_vector<int> device_output_row(gidx.size(), 0); }
auto d_output_row = device_output_row.data(); }
dh::CubMemory temp_memory; }
dh::TransformLbs(
0, &temp_memory, gidx.size(), device_row_ptr.data(), row_ptr.size() - 1,
[=] __device__(int idx, int ridx) { d_output_row[idx] = ridx; });
dh::safe_cuda(cudaDeviceSynchronize());
return device_output_row;
} }
TEST(lbs, Test) { ASSERT_TRUE(test_lbs() == lbs_seg_output); } void SpeedTest() {
int num_rows = 1000000;
int max_row_size = 100;
dh::CubMemory temp_memory;
thrust::host_vector<int> h_row_ptr;
thrust::host_vector<xgboost::bst_uint> h_rows;
CreateTestData(num_rows, max_row_size, &h_row_ptr, &h_rows);
thrust::device_vector<int> row_ptr = h_row_ptr;
thrust::device_vector<int> output_row(h_rows.size());
auto d_output_row = output_row.data();
dh::Timer t;
dh::TransformLbs(
0, &temp_memory, h_rows.size(), dh::raw(row_ptr), row_ptr.size() - 1,
[=] __device__(size_t idx, size_t ridx) { d_output_row[idx] = ridx; });
dh::safe_cuda(cudaDeviceSynchronize());
double time = t.elapsedSeconds();
const int mb_size = 1048576;
size_t size = (sizeof(int) * h_rows.size()) / mb_size;
printf("size: %llumb, time: %fs, bandwidth: %fmb/s\n", size, time,
size / time);
}
void TestLbs() {
srand(17);
dh::CubMemory temp_memory;
std::vector<int> test_rows = {4, 100, 1000};
std::vector<int> test_max_row_sizes = {4, 100, 1300};
for (auto num_rows : test_rows) {
for (auto max_row_size : test_max_row_sizes) {
thrust::host_vector<int> h_row_ptr;
thrust::host_vector<xgboost::bst_uint> h_rows;
CreateTestData(num_rows, max_row_size, &h_row_ptr, &h_rows);
thrust::device_vector<size_t> row_ptr = h_row_ptr;
thrust::device_vector<int> output_row(h_rows.size());
auto d_output_row = output_row.data();
dh::TransformLbs(0, &temp_memory, h_rows.size(), dh::raw(row_ptr),
row_ptr.size() - 1,
[=] __device__(size_t idx, size_t ridx) {
d_output_row[idx] = ridx;
});
dh::safe_cuda(cudaDeviceSynchronize());
ASSERT_TRUE(h_rows == output_row);
}
}
}
TEST(cub_lbs, Test) { TestLbs(); }