Remove hist builder class.

src/common/hist_util.cc

@@ -8,8 +8,8 @@
 
 #include <vector>
 
-#include "../common/common.h"
-#include "column_matrix.h"
+#include "../data/adapter.h"         // for SparsePageAdapterBatch
+#include "../data/gradient_index.h"  // for GHistIndexMatrix
 #include "quantile.h"
 #include "xgboost/base.h"
 #include "xgboost/context.h"  // Context
@@ -24,9 +24,7 @@
   #define PREFETCH_READ_T0(addr) do {} while (0)
 #endif  // defined(XGBOOST_MM_PREFETCH_PRESENT)
 
-namespace xgboost {
-namespace common {
-
+namespace xgboost::common {
 HistogramCuts::HistogramCuts() {
   cut_ptrs_.HostVector().emplace_back(0);
 }
@@ -350,9 +348,8 @@ void BuildHistDispatch(Span<GradientPair const> gpair, const RowSetCollection::E
 }
 
 template <bool any_missing>
-void GHistBuilder::BuildHist(Span<GradientPair const> gpair,
-                             const RowSetCollection::Elem row_indices, const GHistIndexMatrix &gmat,
-                             GHistRow hist, bool force_read_by_column) const {
+void BuildHist(Span<GradientPair const> gpair, const RowSetCollection::Elem row_indices,
+               const GHistIndexMatrix &gmat, GHistRow hist, bool force_read_by_column) {
   /* force_read_by_column is used for testing the columnwise building of histograms.
    * default force_read_by_column = false
    */
@@ -369,14 +366,13 @@ void GHistBuilder::BuildHist(Span<GradientPair const> gpair,
       });
 }
 
-template void GHistBuilder::BuildHist<true>(Span<GradientPair const> gpair,
-                                            const RowSetCollection::Elem row_indices,
-                                            const GHistIndexMatrix &gmat, GHistRow hist,
-                                            bool force_read_by_column) const;
-
-template void GHistBuilder::BuildHist<false>(Span<GradientPair const> gpair,
-                                             const RowSetCollection::Elem row_indices,
-                                             const GHistIndexMatrix &gmat, GHistRow hist,
-                                             bool force_read_by_column) const;
-}  // namespace common
-}  // namespace xgboost
+template void BuildHist<true>(Span<GradientPair const> gpair,
+                              const RowSetCollection::Elem row_indices,
+                              const GHistIndexMatrix &gmat, GHistRow hist,
+                              bool force_read_by_column);
+
+template void BuildHist<false>(Span<GradientPair const> gpair,
+                               const RowSetCollection::Elem row_indices,
+                               const GHistIndexMatrix &gmat, GHistRow hist,
+                               bool force_read_by_column);
+}  // namespace xgboost::common

src/common/hist_util.h

@@ -16,11 +16,9 @@
 #include <vector>
 
 #include "categorical.h"
-#include "common.h"
 #include "quantile.h"
 #include "row_set.h"
 #include "threading_utils.h"
-#include "timer.h"
 #include "xgboost/base.h"  // for bst_feature_t, bst_bin_t
 #include "xgboost/data.h"
 
@@ -598,6 +596,8 @@ class ParallelGHistBuilder {
     }
   }
 
+  [[nodiscard]] bst_bin_t TotalBins() const { return nbins_; }
+
  private:
   void MatchNodeNidPairToHist() {
     size_t hist_allocated_additionally = 0;
@@ -643,27 +643,10 @@ class ParallelGHistBuilder {
   std::map<std::pair<size_t, size_t>, int> tid_nid_to_hist_;
 };
 
-/*!
- * \brief builder for histograms of gradient statistics
- */
-class GHistBuilder {
- public:
-  GHistBuilder() = default;
-  explicit GHistBuilder(uint32_t nbins): nbins_{nbins} {}
-
-  // construct a histogram via histogram aggregation
-  template <bool any_missing>
-  void BuildHist(Span<GradientPair const> gpair, const RowSetCollection::Elem row_indices,
-                 const GHistIndexMatrix& gmat, GHistRow hist,
-                 bool force_read_by_column = false) const;
-  uint32_t GetNumBins() const {
-      return nbins_;
-  }
-
- private:
-  /*! \brief number of all bins over all features */
-  uint32_t nbins_ { 0 };
-};
+// construct a histogram via histogram aggregation
+template <bool any_missing>
+void BuildHist(Span<GradientPair const> gpair, const RowSetCollection::Elem row_indices,
+               const GHistIndexMatrix& gmat, GHistRow hist, bool force_read_by_column = false);
 }  // namespace common
 }  // namespace xgboost
 #endif  // XGBOOST_COMMON_HIST_UTIL_H_

src/common/threading_utils.h

@@ -30,9 +30,7 @@ inline int32_t omp_get_thread_limit() { return std::numeric_limits<int32_t>::max
 }
 #endif  // defined(_MSC_VER)
 
-namespace xgboost {
-namespace common {
-
+namespace xgboost::common {
 // Represent simple range of indexes [begin, end)
 // Inspired by tbb::blocked_range
 class Range1d {
@@ -69,7 +67,7 @@ class Range1d {
 // [1,2], [3,4], [5,6], [7,8], [9]
 // The class helps to process data in several tree nodes (non-balanced usually) in parallel
 // Using nested parallelism (by nodes and by data in each node)
-// it helps  to improve CPU resources utilization
+// it helps to improve CPU resources utilization
 class BlockedSpace2d {
  public:
   // Example of space:
@@ -86,39 +84,47 @@ class BlockedSpace2d {
   // dim1 - size of the first dimension in the space
   // getter_size_dim2 - functor to get the second dimensions for each 'row' by row-index
   // grain_size - max size of produced blocks
-  template<typename Func>
-  BlockedSpace2d(size_t dim1, Func getter_size_dim2, size_t grain_size) {
-    for (size_t i = 0; i < dim1; ++i) {
-      const size_t size = getter_size_dim2(i);
-      const size_t n_blocks = size/grain_size + !!(size % grain_size);
-      for (size_t iblock = 0; iblock < n_blocks; ++iblock) {
-        const size_t begin = iblock * grain_size;
-        const size_t end   = std::min(begin + grain_size, size);
+  template <typename Func>
+  BlockedSpace2d(std::size_t dim1, Func getter_size_dim2, std::size_t grain_size) {
+    for (std::size_t i = 0; i < dim1; ++i) {
+      std::size_t size = getter_size_dim2(i);
+      // Each row (second dim) is divided into n_blocks
+      std::size_t n_blocks = size / grain_size + !!(size % grain_size);
+      for (std::size_t iblock = 0; iblock < n_blocks; ++iblock) {
+        std::size_t begin = iblock * grain_size;
+        std::size_t end = std::min(begin + grain_size, size);
         AddBlock(i, begin, end);
       }
     }
   }
 
   // Amount of blocks(tasks) in a space
-  size_t Size() const {
+  [[nodiscard]] std::size_t Size() const {
     return ranges_.size();
   }
 
   // get index of the first dimension of i-th block(task)
-  size_t GetFirstDimension(size_t i) const {
+  [[nodiscard]] std::size_t GetFirstDimension(size_t i) const {
     CHECK_LT(i, first_dimension_.size());
     return first_dimension_[i];
   }
 
   // get a range of indexes for the second dimension of i-th block(task)
-  Range1d GetRange(size_t i) const {
+  [[nodiscard]] Range1d GetRange(size_t i) const {
     CHECK_LT(i, ranges_.size());
     return ranges_[i];
   }
 
  private:
-  void AddBlock(size_t first_dimension, size_t begin, size_t end) {
-    first_dimension_.push_back(first_dimension);
+  /**
+   * @brief Add a parallel block.
+   *
+   * @param first_dim The row index.
+   * @param begin     The begin of the second dimension.
+   * @param end       The end of the second dimension.
+   */
+  void AddBlock(std::size_t first_dim, std::size_t begin, std::size_t end) {
+    first_dimension_.push_back(first_dim);
     ranges_.emplace_back(begin, end);
   }
 
@@ -303,7 +309,6 @@ class MemStackAllocator {
  * \brief Constant that can be used for initializing static thread local memory.
  */
 std::int32_t constexpr DefaultMaxThreads() { return 128; }
-}  // namespace common
-}  // namespace xgboost
+}  // namespace xgboost::common
 
 #endif  // XGBOOST_COMMON_THREADING_UTILS_H_

src/tree/hist/histogram.h

@@ -22,7 +22,6 @@ class HistogramBuilder {
   common::HistCollection hist_;
   /*! \brief culmulative local parent histogram of gradients. */
   common::HistCollection hist_local_worker_;
-  common::GHistBuilder builder_;
   common::ParallelGHistBuilder buffer_;
   BatchParam param_;
   int32_t n_threads_{-1};
@@ -49,7 +48,6 @@ class HistogramBuilder {
     hist_.Init(total_bins);
     hist_local_worker_.Init(total_bins);
     buffer_.Init(total_bins);
-    builder_ = common::GHistBuilder(total_bins);
     is_distributed_ = is_distributed;
     is_col_split_ = is_col_split;
     // Workaround s390x gcc 7.5.0
@@ -88,8 +86,7 @@ class HistogramBuilder {
                                                     elem.begin + end_of_row_set, nid);
       auto hist = buffer_.GetInitializedHist(tid, nid_in_set);
       if (rid_set.Size() != 0) {
-        builder_.template BuildHist<any_missing>(gpair_h, rid_set, gidx, hist,
-                                                 force_read_by_column);
+        common::BuildHist<any_missing>(gpair_h, rid_set, gidx, hist, force_read_by_column);
       }
     });
   }
@@ -163,9 +160,9 @@ class HistogramBuilder {
                                 std::vector<ExpandEntry> const &nodes_for_explicit_hist_build,
                                 std::vector<ExpandEntry> const &nodes_for_subtraction_trick,
                                 int starting_index, int sync_count) {
-    const size_t nbins = builder_.GetNumBins();
+    auto n_bins = buffer_.TotalBins();
     common::BlockedSpace2d space(
-        nodes_for_explicit_hist_build.size(), [&](size_t) { return nbins; }, 1024);
+        nodes_for_explicit_hist_build.size(), [&](size_t) { return n_bins; }, 1024);
     common::ParallelFor2d(space, n_threads_, [&](size_t node, common::Range1d r) {
       const auto &entry = nodes_for_explicit_hist_build[node];
       auto this_hist = this->hist_[entry.nid];
@@ -188,22 +185,21 @@ class HistogramBuilder {
     });
 
     collective::Allreduce<collective::Operation::kSum>(
-        reinterpret_cast<double *>(this->hist_[starting_index].data()),
-        builder_.GetNumBins() * sync_count * 2);
+        reinterpret_cast<double *>(this->hist_[starting_index].data()), n_bins * sync_count * 2);
 
     ParallelSubtractionHist(space, nodes_for_explicit_hist_build, nodes_for_subtraction_trick,
                             p_tree);
 
     common::BlockedSpace2d space2(
-        nodes_for_subtraction_trick.size(), [&](size_t) { return nbins; }, 1024);
+        nodes_for_subtraction_trick.size(), [&](size_t) { return n_bins; }, 1024);
     ParallelSubtractionHist(space2, nodes_for_subtraction_trick, nodes_for_explicit_hist_build,
                             p_tree);
   }
 
   void SyncHistogramLocal(RegTree const *p_tree,
                           std::vector<ExpandEntry> const &nodes_for_explicit_hist_build,
                           std::vector<ExpandEntry> const &nodes_for_subtraction_trick) {
-    const size_t nbins = this->builder_.GetNumBins();
+    const size_t nbins = this->buffer_.TotalBins();
     common::BlockedSpace2d space(
         nodes_for_explicit_hist_build.size(), [&](size_t) { return nbins; }, 1024);
 

tests/cpp/tree/hist/test_evaluate_splits.cc

@@ -48,12 +48,10 @@ void TestEvaluateSplits(bool force_read_by_column) {
   std::iota(row_indices.begin(), row_indices.end(), 0);
   row_set_collection.Init();
 
-  auto hist_builder = common::GHistBuilder(gmat.cut.Ptrs().back());
   hist.Init(gmat.cut.Ptrs().back());
   hist.AddHistRow(0);
   hist.AllocateAllData();
-  hist_builder.template BuildHist<false>(row_gpairs, row_set_collection[0],
-                                         gmat, hist[0], force_read_by_column);
+  common::BuildHist<false>(row_gpairs, row_set_collection[0], gmat, hist[0], force_read_by_column);
 
   // Compute total gradient for all data points
   GradientPairPrecise total_gpair;

tests/cpp/tree/hist/test_histogram.cc

@@ -13,8 +13,7 @@
 #include "../../categorical_helpers.h"
 #include "../../helpers.h"
 
-namespace xgboost {
-namespace tree {
+namespace xgboost::tree {
 namespace {
 void InitRowPartitionForTest(common::RowSetCollection *row_set, size_t n_samples, size_t base_rowid = 0) {
   auto &row_indices = *row_set->Data();
@@ -487,5 +486,5 @@ TEST(CPUHistogram, ExternalMemory) {
   TestHistogramExternalMemory(&ctx, {kBins, sparse_thresh}, false, false);
   TestHistogramExternalMemory(&ctx, {kBins, sparse_thresh}, false, true);
 }
-}  // namespace tree
-}  // namespace xgboost
+}  // namespace xgboost::tree
+