debug.

Author: trivialfis

src/common/io.cc

@@ -212,7 +212,8 @@ char* PrivateMmapStream::Open(std::string path, bool read_only, std::size_t offs
   char* ptr{nullptr};
   auto view_start = offset / GetPageSize() * GetPageSize();
   auto view_size = length + (offset - view_start);
-  std::cout << view_start << " size: " << view_size << std::endl;
+  std::cout << "offset:" << offset << ", length:" << length << ", start:" << view_start << ", size:" << view_size << std::endl;
+  // std::cout << view_start << " size: " << view_size << std::endl;
 #if defined(__linux__) || defined(__GLIBC__)
   int prot{PROT_READ};
   if (!read_only) {

src/data/sparse_page_raw_format.cc

@@ -22,6 +22,7 @@ class SparsePageRawFormat : public SparsePageFormat<T> {
     if (!fi->Read(&offset_vec)) {
       return false;
     }
+    std::cout << "read size:" << offset_vec.size() << " back:" << offset_vec.back() << std::endl;
     auto& data_vec = page->data.HostVector();
     CHECK_NE(page->offset.Size(), 0U) << "Invalid SparsePage file";
     data_vec.resize(offset_vec.back());
@@ -37,6 +38,7 @@ class SparsePageRawFormat : public SparsePageFormat<T> {
 
   size_t Write(const T& page, dmlc::Stream* fo) override {
     const auto& offset_vec = page.offset.HostVector();
+    std::cout << "write size:" << offset_vec.size() << " back:" << offset_vec.back() << std::endl;
     const auto& data_vec = page.data.HostVector();
     CHECK(page.offset.Size() != 0 && offset_vec[0] == 0);
     CHECK_EQ(offset_vec.back(), page.data.Size());

src/data/sparse_page_source.h

@@ -115,7 +115,7 @@ class SparsePageSourceImpl : public BatchIteratorImpl<S> {
     }
     // An heuristic for number of pre-fetched batches.  We can make it part of BatchParam
     // to let user adjust number of pre-fetched batches when needed.
-    uint32_t constexpr kPreFetch = 4;
+    uint32_t constexpr kPreFetch = 1;
 
     size_t n_prefetch_batches = std::min(kPreFetch, n_batches_);
     CHECK_GT(n_prefetch_batches, 0) << "total batches:" << n_batches_;
@@ -132,7 +132,7 @@ class SparsePageSourceImpl : public BatchIteratorImpl<S> {
       CHECK_LT(fetch_it, cache_info_->offset.size());
       ring_->at(fetch_it) = std::async(std::launch::async, [fetch_it, self, this]() {
         auto page = std::make_shared<S>();
-        this->exec_.Run([&] {
+//        this->exec_.Run([&] {
           common::Timer timer;
           timer.Start();
           std::unique_ptr<SparsePageFormat<S>> fmt{CreatePageFormat<S>("raw")};
@@ -144,7 +144,7 @@ class SparsePageSourceImpl : public BatchIteratorImpl<S> {
           auto fi = std::make_unique<common::PrivateMmapStream>(n, true, offset, length);
           CHECK(fmt->Read(page.get(), fi.get()));
           LOG(INFO) << "Read a page in " << timer.ElapsedSeconds() << " seconds.";
-        });
+ //       });
         return page;
       });
     }
@@ -174,6 +174,7 @@ class SparsePageSourceImpl : public BatchIteratorImpl<S> {
     }
 
     auto bytes = fmt->Write(*page_, fo.get());
+    std::cout << "wrote: " << bytes << std::endl;
 
     timer.Stop();
     LOG(INFO) << static_cast<double>(bytes) / 1024.0 / 1024.0 << " MB written in "

tests/cpp/common/test_io.cc

@@ -103,7 +103,7 @@ TEST(IO, PrivateMmapStream) {
   {
     std::unique_ptr<dmlc::Stream> fo{dmlc::Stream::Create(path.c_str(), "w")};
     for (std::size_t i = 0; i < n_batches; ++i) {
-      std::size_t size = (i + 1) * 2;
+      std::size_t size = (i + 1) * 8192;
       std::vector<T> data(size, 0);
       std::iota(data.begin(), data.end(), i * i);
 

tests/cpp/test_main.cc

@@ -11,7 +11,8 @@
 int main(int argc, char ** argv) {
   xgboost::Args args {{"verbosity", "2"}};
   xgboost::ConsoleLogger::Configure(args);
-
+  ::testing::GTEST_FLAG(filter) = "CPUHistogram.ExternalMemory*";
+  // ::testing::GTEST_FLAG(filter) = "IO.PrivateMmapStream";
   testing::InitGoogleTest(&argc, argv);
   testing::FLAGS_gtest_death_test_style = "threadsafe";
   auto rmm_alloc = xgboost::SetUpRMMResourceForCppTests(argc, argv);

tests/cpp/tree/hist/test_histogram.cc

@@ -477,15 +477,8 @@ TEST(CPUHistogram, ExternalMemory) {
   int32_t constexpr kBins = 256;
   Context ctx;
 
+  std::cout << "l:" << __LINE__ << std::endl;
   TestHistogramExternalMemory(&ctx, BatchParam{kBins, common::Span<float>{}, false}, true, false);
-  TestHistogramExternalMemory(&ctx, BatchParam{kBins, common::Span<float>{}, false}, true, true);
-
-  float sparse_thresh{0.5};
-  TestHistogramExternalMemory(&ctx, {kBins, sparse_thresh}, false, false);
-  TestHistogramExternalMemory(&ctx, {kBins, sparse_thresh}, false, true);
-  sparse_thresh = std::numeric_limits<float>::quiet_NaN();
-  TestHistogramExternalMemory(&ctx, {kBins, sparse_thresh}, false, false);
-  TestHistogramExternalMemory(&ctx, {kBins, sparse_thresh}, false, true);
 }
 }  // namespace tree
 }  // namespace xgboost