Skip to content

[EM] Prevent blocking memcpy during prediction. #11548

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 8, 2025
Merged

[EM] Prevent blocking memcpy during prediction. #11548

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions include/xgboost/host_device_vector.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -101,6 +101,7 @@ class HostDeviceVector {

[[nodiscard]] bool Empty() const { return Size() == 0; }
[[nodiscard]] std::size_t Size() const;
[[nodiscard]] std::size_t SizeBytes() const { return this->Size() * sizeof(T); }
[[nodiscard]] DeviceOrd Device() const;
common::Span<T> DeviceSpan();
common::Span<const T> ConstDeviceSpan() const;
Expand Down
17 changes: 17 additions & 0 deletions src/predictor/gpu_predictor.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -364,6 +364,20 @@ class DeviceModel {
int num_group;
CatContainer const* cat_enc{nullptr};

[[nodiscard]] std::size_t MemCostBytes() const {
std::size_t n_bytes = 0;
n_bytes += stats.ConstDeviceSpan().size_bytes();
n_bytes += tree_segments.ConstDeviceSpan().size_bytes();
n_bytes += nodes.ConstDeviceSpan().size_bytes();
n_bytes += tree_group.ConstDeviceSpan().size_bytes();
n_bytes += split_types.ConstDeviceSpan().size_bytes();
n_bytes += categories_tree_segments.ConstDeviceSpan().size_bytes();
n_bytes += categories_node_segments.ConstDeviceSpan().size_bytes();
n_bytes += categories.ConstDeviceSpan().size_bytes();
n_bytes += sizeof(tree_beg_) + sizeof(tree_end_) + sizeof(num_group) + sizeof(cat_enc);
return n_bytes;
}

void Init(const gbm::GBTreeModel& model, bst_tree_t tree_begin, bst_tree_t tree_end,
DeviceOrd device) {
dh::safe_cuda(cudaSetDevice(device.ordinal));
Expand Down Expand Up @@ -438,6 +452,9 @@ class DeviceModel {

this->cat_enc = model.Cats();
CHECK(this->cat_enc);

auto n_bytes = this->MemCostBytes(); // Pull data to device, and get the size of the model.
LOG(DEBUG) << "Model size:" << common::HumanMemUnit(n_bytes);
}
};

Expand Down
Loading