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

Implement unified update prediction cache for (gpu_)hist. (#6860)

Browse Source 
* Implement utilites for linalg.
* Unify the update prediction cache functions.
* Implement update prediction cache for multi-class gpu hist.
This commit is contained in:
Jiaming Yuan
2021-04-17 00:29:34 +08:00
committed by GitHub
parent 1b26a2a561
commit 556a83022d
10 changed files with 246 additions and 68 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
tests/cpp/test_learner.cc
View File

@@ -294,6 +294,13 @@ TEST(Learner, GPUConfiguration) {
learner->UpdateOneIter(0, p_dmat);
ASSERT_EQ(learner->GetGenericParameter().gpu_id, 0);
}
{
std::unique_ptr<Learner> learner {Learner::Create(mat)};
learner->SetParams({Arg{"tree_method", "gpu_hist"},
Arg{"gpu_id", "-1"}});
learner->UpdateOneIter(0, p_dmat);
ASSERT_EQ(learner->GetGenericParameter().gpu_id, 0);
}
{
// with CPU algorithm
std::unique_ptr<Learner> learner {Learner::Create(mat)};