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

Improve test coverage with predictor configuration. (#9354)

Browse Source 
* Improve test coverage with predictor configuration.

- Test with ext memory.
- Test with QDM.
- Test with dart.
This commit is contained in:
Jiaming Yuan
2023-07-05 15:17:22 +08:00
committed by GitHub
parent 6c9c8a9001
commit 645037e376
17 changed files with 280 additions and 79 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
tests/cpp/data/test_gradient_index.cc
View File

@@ -27,26 +27,31 @@
#include "xgboost/host_device_vector.h" // for HostDeviceVector
namespace xgboost::data {
TEST(GradientIndex, ExternalMemory) {
TEST(GradientIndex, ExternalMemoryBaseRowID) {
Context ctx;
std::unique_ptr<DMatrix> dmat = CreateSparsePageDMatrix(10000);
auto p_fmat = RandomDataGenerator{4096, 256, 0.5}
.Device(ctx.gpu_id)
.Batches(8)
.GenerateSparsePageDMatrix("cache", true);
std::vector<size_t> base_rowids;
std::vector<float> hessian(dmat->Info().num_row_, 1);
for (auto const &page : dmat->GetBatches<GHistIndexMatrix>(&ctx, {64, hessian, true})) {
std::vector<float> hessian(p_fmat->Info().num_row_, 1);
for (auto const &page : p_fmat->GetBatches<GHistIndexMatrix>(&ctx, {64, hessian, true})) {
base_rowids.push_back(page.base_rowid);
}
size_t i = 0;
for (auto const &page : dmat->GetBatches<SparsePage>()) {
std::size_t i = 0;
for (auto const &page : p_fmat->GetBatches<SparsePage>()) {
ASSERT_EQ(base_rowids[i], page.base_rowid);
++i;
}
base_rowids.clear();
for (auto const &page : dmat->GetBatches<GHistIndexMatrix>(&ctx, {64, hessian, false})) {
for (auto const &page : p_fmat->GetBatches<GHistIndexMatrix>(&ctx, {64, hessian, false})) {
base_rowids.push_back(page.base_rowid);
}
i = 0;
for (auto const &page : dmat->GetBatches<SparsePage>()) {
for (auto const &page : p_fmat->GetBatches<SparsePage>()) {
ASSERT_EQ(base_rowids[i], page.base_rowid);
++i;
}

10
tests/cpp/data/test_sparse_page_dmatrix.cc
View File

@@ -76,9 +76,11 @@ TEST(SparsePageDMatrix, LoadFile) {
// allow caller to retain pages so they can process multiple pages at the same time.
template <typename Page>
void TestRetainPage() {
auto m = CreateSparsePageDMatrix(10000);
std::size_t n_batches = 4;
auto p_fmat = RandomDataGenerator{1024, 128, 0.5f}.Batches(n_batches).GenerateSparsePageDMatrix(
"cache", true);
Context ctx;
auto batches = m->GetBatches<Page>(&ctx);
auto batches = p_fmat->GetBatches<Page>(&ctx);
auto begin = batches.begin();
auto end = batches.end();
@@ -94,7 +96,7 @@ void TestRetainPage() {
}
ASSERT_EQ(pages.back().Size(), (*it).Size());
}
ASSERT_GE(iterators.size(), 2);
ASSERT_GE(iterators.size(), n_batches);
for (size_t i = 0; i < iterators.size(); ++i) {
ASSERT_EQ((*iterators[i]).Size(), pages.at(i).Size());
@@ -102,7 +104,7 @@ void TestRetainPage() {
}
// make sure it's const and the caller can not modify the content of page.
for (auto &page : m->GetBatches<Page>({&ctx})) {
for (auto &page : p_fmat->GetBatches<Page>({&ctx})) {
static_assert(std::is_const<std::remove_reference_t<decltype(page)>>::value);
}
}