Skip to content

Fix the pointer_residency_count function #2887

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

Closed
viclafargue wants to merge 1 commit into from
Closed

Fix the pointer_residency_count function #2887

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
16 changes: 13 additions & 3 deletions cpp/include/raft/spatial/knn/detail/ann_utils.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
/*
* SPDX-FileCopyrightText: Copyright (c) 2022-2024, NVIDIA CORPORATION.
* SPDX-FileCopyrightText: Copyright (c) 2022-2025, NVIDIA CORPORATION.
* SPDX-License-Identifier: Apache-2.0
*/

Expand Down Expand Up @@ -50,8 +50,18 @@ struct pointer_residency_count<Type, Types...> {
auto [on_device, on_host] = pointer_residency_count<Types...>::run(ptrs...);
cudaPointerAttributes attr;
RAFT_CUDA_TRY(cudaPointerGetAttributes(&attr, ptr));
if (attr.devicePointer || attr.type == cudaMemoryTypeDevice) { ++on_device; }
if (attr.hostPointer || attr.type == cudaMemoryTypeUnregistered) { ++on_host; }

// Unregistered memory (e.g. regular malloc, numpy arrays) is host-only.
// With HMM/ATS enabled, cudaPointerGetAttributes may report a non-null devicePointer
// for such memory, but relying on page-fault access is unreliable for GPU kernels.
// We must treat unregistered memory as host-only to ensure proper device allocation.
bool is_unregistered = (attr.type == cudaMemoryTypeUnregistered);
bool is_device_accessible =
!is_unregistered && (attr.devicePointer || attr.type == cudaMemoryTypeDevice);
bool is_host_accessible = is_unregistered || attr.hostPointer;

if (is_device_accessible) { ++on_device; }
if (is_host_accessible) { ++on_host; }
return std::make_tuple(on_device, on_host);
}
};
Expand Down