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

Optimize ‘hist’ for multi-core CPU (#4529)

Browse Source 
* Initial performance optimizations for xgboost

* remove includes

* revert float->double

* fix for CI

* fix for CI

* fix for CI

* fix for CI

* fix for CI

* fix for CI

* fix for CI

* fix for CI

* fix for CI

* fix for CI

* Check existence of _mm_prefetch and __builtin_prefetch

* Fix lint

* optimizations for CPU

* appling comments in review

* add some comments, code refactoring

* fixing issues in CI

* adding runtime checks

* remove 1 extra check

* remove extra checks in BuildHist

* remove checks

* add debug info

* added debug info

* revert changes

* added comments

* Apply suggestions from code review

Co-Authored-By: Philip Hyunsu Cho <chohyu01@cs.washington.edu>

* apply review comments

* Remove unused function CreateNewNodes()

* Add descriptive comment on node_idx variable in QuantileHistMaker::Builder::BuildHistsBatch()
This commit is contained in:
Egor Smirnov
2019-06-27 22:33:49 +04:00
committed by Philip Hyunsu Cho
parent abffbe014e
commit 4d6590be3c
9 changed files with 1342 additions and 818 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
tests/cpp/tree/test_quantile_hist.cc
View File

@@ -101,8 +101,13 @@ class QuantileHistMock : public QuantileHistMaker {
RealImpl::InitData(gmat, gpair, fmat, tree);
GHistIndexBlockMatrix dummy;
hist_.AddHistRow(nid);
BuildHist(gpair, row_set_collection_[nid],
gmat, dummy, hist_[nid], false);
std::vector<std::vector<float*>> hist_buffers;
std::vector<std::vector<uint8_t>> hist_is_init;
std::vector<ExpandEntry> nodes = {ExpandEntry(nid, -1, -1, tree.GetDepth(0), 0.0, 0)};
BuildHistsBatch(nodes, const_cast<RegTree*>(&tree), gmat, gpair, &hist_buffers, &hist_is_init);
RealImpl::InitNewNode(nid, gmat, gpair, fmat, const_cast<RegTree*>(&tree), &snode_[0], tree[0].Parent());
EvaluateSplitsBatch(nodes, gmat, fmat, hist_is_init, hist_buffers);
// Check if number of histogram bins is correct
ASSERT_EQ(hist_[nid].size(), gmat.cut.row_ptr.back());
@@ -143,10 +148,12 @@ class QuantileHistMock : public QuantileHistMaker {
RealImpl::InitData(gmat, row_gpairs, *(*dmat), tree);
hist_.AddHistRow(0);
BuildHist(row_gpairs, row_set_collection_[0],
gmat, quantile_index_block, hist_[0], false);
RealImpl::InitNewNode(0, gmat, row_gpairs, *(*dmat), tree);
std::vector<ExpandEntry> nodes = {ExpandEntry(0, -1, -1, tree.GetDepth(0), 0.0, 0)};
std::vector<std::vector<float*>> hist_buffers;
std::vector<std::vector<uint8_t>> hist_is_init;
BuildHistsBatch(nodes, const_cast<RegTree*>(&tree), gmat, row_gpairs, &hist_buffers, &hist_is_init);
RealImpl::InitNewNode(0, gmat, row_gpairs, *(*dmat), const_cast<RegTree*>(&tree), &snode_[0], tree[0].Parent());
EvaluateSplitsBatch(nodes, gmat, **dmat, hist_is_init, hist_buffers);
/* Compute correct split (best_split) using the computed histogram */
const size_t num_row = dmat->get()->Info().num_row_;
@@ -197,6 +204,7 @@ class QuantileHistMock : public QuantileHistMaker {
const auto split_gain
= evaluator->ComputeSplitScore(0, fid, GradStats(left_sum),
GradStats(right_sum));
if (split_gain > best_split_gain) {
best_split_gain = split_gain;
best_split_feature = fid;
@@ -206,7 +214,8 @@ class QuantileHistMock : public QuantileHistMaker {
}
/* Now compare against result given by EvaluateSplit() */
RealImpl::EvaluateSplit(0, gmat, hist_, *(*dmat), tree);
EvaluateSplitsBatch(nodes, gmat, **dmat, hist_is_init, hist_buffers);
ASSERT_EQ(snode_[0].best.SplitIndex(), best_split_feature);
ASSERT_EQ(snode_[0].best.split_value, gmat.cut.cut[best_split_threshold]);
@@ -289,7 +298,7 @@ TEST(Updater, QuantileHist_EvalSplits) {
std::vector<std::pair<std::string, std::string>> cfg
{{"num_feature", std::to_string(QuantileHistMock::GetNumColumns())},
{"split_evaluator", "elastic_net"},
{"reg_lambda", "0"}, {"reg_alpha", "0"}, {"max_delta_step", "0"},
{"reg_lambda", "1.0f"}, {"reg_alpha", "0"}, {"max_delta_step", "0"},
{"min_child_weight", "0"}};
QuantileHistMock maker(cfg);
maker.TestEvaluateSplit();