From 033a666900a2b58f72a3050e63b02f1681e4f609 Mon Sep 17 00:00:00 2001
From: Jiaming Yuan <jm.yuan@outlook.com>
Date: Sat, 17 Aug 2024 01:35:47 +0800
Subject: [PATCH] [EM] Log the page size of ellpack. (#10713)

---
 src/common/timer.h              | 14 +++++---------
 src/data/ellpack_page.cu        | 11 +----------
 src/data/ellpack_page.cuh       |  3 +--
 src/data/ellpack_page_source.cu |  2 ++
 4 files changed, 9 insertions(+), 21 deletions(-)

diff --git a/src/common/timer.h b/src/common/timer.h
index 80748e7b8264..3308dfb653dd 100644
--- a/src/common/timer.h
+++ b/src/common/timer.h
@@ -1,18 +1,15 @@
-/*!
- * Copyright by Contributors 2017-2019
+/**
+ * Copyright 2017-2024, XGBoost Contributors
  */
 #pragma once
 #include <xgboost/logging.h>
+
 #include <chrono>
-#include <iostream>
 #include <map>
 #include <string>
 #include <utility>
-#include <vector>
-
-namespace xgboost {
-namespace common {
 
+namespace xgboost::common {
 struct Timer {
   using ClockT = std::chrono::high_resolution_clock;
   using TimePointT = std::chrono::high_resolution_clock::time_point;
@@ -82,5 +79,4 @@ struct Monitor {
   void Start(const std::string &name);
   void Stop(const std::string &name);
 };
-}  // namespace common
-}  // namespace xgboost
+}  // namespace xgboost::common
diff --git a/src/data/ellpack_page.cu b/src/data/ellpack_page.cu
index 575bcd5ce9a5..b7ec72ad393c 100644
--- a/src/data/ellpack_page.cu
+++ b/src/data/ellpack_page.cu
@@ -356,7 +356,6 @@ EllpackPageImpl::EllpackPageImpl(Context const* ctx, GHistIndexMatrix const& pag
     : is_dense{page.IsDense()},
       base_rowid{page.base_rowid},
       n_rows{page.Size()},
-      // This makes a copy of the cut values.
       cuts_{std::make_shared<common::HistogramCuts>(page.cut)} {
   auto it = common::MakeIndexTransformIter(
       [&](size_t i) { return page.row_ptr[i + 1] - page.row_ptr[i]; });
@@ -540,15 +539,7 @@ void EllpackPageImpl::CreateHistIndices(DeviceOrd device,
 // Return the number of rows contained in this page.
 [[nodiscard]] bst_idx_t EllpackPageImpl::Size() const { return n_rows; }
 
-// Return the memory cost for storing the compressed features.
-size_t EllpackPageImpl::MemCostBytes(size_t num_rows, size_t row_stride,
-                                     const common::HistogramCuts& cuts) {
-  // Required buffer size for storing data matrix in EtoLLPack format.
-  size_t compressed_size_bytes =
-      common::CompressedBufferWriter::CalculateBufferSize(row_stride * num_rows,
-                                                          cuts.TotalBins() + 1);
-  return compressed_size_bytes;
-}
+std::size_t EllpackPageImpl::MemCostBytes() const { return this->gidx_buffer.size_bytes(); }
 
 EllpackDeviceAccessor EllpackPageImpl::GetDeviceAccessor(
     DeviceOrd device, common::Span<FeatureType const> feature_types) const {
diff --git a/src/data/ellpack_page.cuh b/src/data/ellpack_page.cuh
index f11bdfae1d97..af11dec3fc15 100644
--- a/src/data/ellpack_page.cuh
+++ b/src/data/ellpack_page.cuh
@@ -217,8 +217,7 @@ class EllpackPageImpl {
 
   [[nodiscard]] bool IsDense() const { return is_dense; }
   /** @return Estimation of memory cost of this page. */
-  static size_t MemCostBytes(size_t num_rows, size_t row_stride, const common::HistogramCuts&cuts) ;
-
+  std::size_t MemCostBytes() const;
 
   /**
    * @brief Return the total number of symbols (total number of bins plus 1 for not
diff --git a/src/data/ellpack_page_source.cu b/src/data/ellpack_page_source.cu
index 7ab4819e14e1..2927d028cf79 100644
--- a/src/data/ellpack_page_source.cu
+++ b/src/data/ellpack_page_source.cu
@@ -172,6 +172,8 @@ void EllpackPageSourceImpl<F>::Fetch() {
     Context ctx = Context{}.MakeCUDA(this->Device().ordinal);
     *impl = EllpackPageImpl{&ctx, this->GetCuts(), *csr, is_dense_, row_stride_, feature_types_};
     this->page_->SetBaseRowId(csr->base_rowid);
+    LOG(INFO) << "Generated an Ellpack page with size: " << impl->MemCostBytes()
+              << " from a SparsePage with size:" << csr->MemCostBytes();
     this->WriteCache();
   }
 }