Fix external memory for get column batches. (#4622)

* Fix external memory for get column batches.

This fixes two bugs:

* Use PushCSC for get column batches.
* Don't remove the created temporary directory before finishing test.

* Check all pages.
This commit is contained in:
Jiaming Yuan
2019-06-30 09:56:49 +08:00
committed by GitHub
parent a30176907f
commit 45876bf41b
14 changed files with 90 additions and 46 deletions

View File

@@ -412,6 +412,18 @@ void SparsePage::PushCSC(const SparsePage &batch) {
self_offset = std::move(offset);
}
void SparsePage::Push(const Inst &inst) {
auto& data_vec = data.HostVector();
auto& offset_vec = offset.HostVector();
offset_vec.push_back(offset_vec.back() + inst.size());
size_t begin = data_vec.size();
data_vec.resize(begin + inst.size());
if (inst.size() != 0) {
std::memcpy(dmlc::BeginPtr(data_vec) + begin, inst.data(),
sizeof(Entry) * inst.size());
}
}
namespace data {
// List of files that will be force linked in static links.
DMLC_REGISTRY_LINK_TAG(sparse_page_raw_format);

View File

@@ -20,7 +20,7 @@
namespace xgboost {
namespace data {
// Used for single batch data.
class SimpleDMatrix : public DMatrix {
public:
explicit SimpleDMatrix(std::unique_ptr<DataSource>&& source)

View File

@@ -18,7 +18,7 @@
namespace xgboost {
namespace data {
// Used for external memory.
class SparsePageDMatrix : public DMatrix {
public:
explicit SparsePageDMatrix(std::unique_ptr<DataSource>&& source,

View File

@@ -221,8 +221,8 @@ void SparsePageSource::CreateRowPage(dmlc::Parser<uint32_t>* src,
CHECK(info.qids_.empty() || info.qids_.size() == info.num_row_);
info.SaveBinary(fo.get());
}
LOG(CONSOLE) << "SparsePageSource::CreateRowPage Finished writing to "
<< name_info;
LOG(INFO) << "SparsePageSource::CreateRowPage Finished writing to "
<< name_info;
}
void SparsePageSource::CreatePageFromDMatrix(DMatrix* src,
@@ -251,7 +251,7 @@ void SparsePageSource::CreatePageFromDMatrix(DMatrix* src,
if (page_type == ".row.page") {
page->Push(batch);
} else if (page_type == ".col.page") {
page->Push(batch.GetTranspose(src->Info().num_col_));
page->PushCSC(batch.GetTranspose(src->Info().num_col_));
} else if (page_type == ".sorted.col.page") {
SparsePage tmp = batch.GetTranspose(src->Info().num_col_);
page->PushCSC(tmp);
@@ -266,9 +266,9 @@ void SparsePageSource::CreatePageFromDMatrix(DMatrix* src,
writer.Alloc(&page);
page->Clear();
double tdiff = dmlc::GetTime() - tstart;
LOG(CONSOLE) << "Writing to " << cache_info << " in "
<< ((bytes_write >> 20UL) / tdiff) << " MB/s, "
<< (bytes_write >> 20UL) << " written";
LOG(INFO) << "Writing to " << cache_info << " in "
<< ((bytes_write >> 20UL) / tdiff) << " MB/s, "
<< (bytes_write >> 20UL) << " written";
}
}
if (page->data.Size() != 0) {
@@ -281,7 +281,7 @@ void SparsePageSource::CreatePageFromDMatrix(DMatrix* src,
fo->Write(&tmagic, sizeof(tmagic));
info.SaveBinary(fo.get());
}
LOG(CONSOLE) << "SparsePageSource: Finished writing to " << name_info;
LOG(INFO) << "SparsePageSource: Finished writing to " << name_info;
}
void SparsePageSource::CreateRowPage(DMatrix* src,

View File

@@ -39,7 +39,7 @@ SparsePageWriter::SparsePageWriter(
qrecycle_.Push(std::move(page));
}
fo.reset(nullptr);
LOG(CONSOLE) << "SparsePage::Writer Finished writing to " << name_shard;
LOG(INFO) << "SparsePage::Writer Finished writing to " << name_shard;
}));
}
}