Skip to content

Revert PR #580 streaming workaround (CCCL #1422 resolved) #810

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
PointKernel merged 2 commits into from
May 7, 2026
Merged

Revert PR #580 streaming workaround (CCCL #1422 resolved) #810

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7e00710
Revert PR #580 streaming workaround (CCCL #1422 resolved)
PointKernel May 1, 2026
c950496
Merge branch 'dev' into revert-pr-580-streaming
PointKernel May 5, 2026
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
80 changes: 34 additions & 46 deletions include/cuco/detail/open_addressing/open_addressing_impl.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -830,60 +830,48 @@ class open_addressing_impl : private open_addressing_compatible<Key, Value, Prob
using temp_allocator_type =
typename std::allocator_traits<allocator_type>::template rebind_alloc<char>;

cuco::detail::index_type constexpr stride = std::numeric_limits<std::int32_t>::max();

cuco::detail::index_type h_num_out{0};
auto temp_allocator = temp_allocator_type{this->allocator()};
auto d_num_out =
reinterpret_cast<size_type*>(temp_allocator.allocate(sizeof(size_type), stream));

// TODO: PR #580 to be reverted once https://github.com/NVIDIA/cccl/issues/1422 is resolved
for (cuco::detail::index_type offset = 0;
offset < static_cast<cuco::detail::index_type>(this->capacity());
offset += stride) {
auto const num_items =
std::min(static_cast<cuco::detail::index_type>(this->capacity()) - offset, stride);
auto const begin = cuda::make_transform_iterator(
cuda::counting_iterator{static_cast<size_type>(offset)},
detail::open_addressing_ns::get_slot<has_payload, storage_ref_type>(this->storage_ref()));
auto const is_filled = detail::open_addressing_ns::slot_is_filled<has_payload, key_type>{
this->empty_key_sentinel(), this->erased_key_sentinel()};

std::size_t temp_storage_bytes = 0;

CUCO_CUDA_TRY(cub::DeviceSelect::If(nullptr,
temp_storage_bytes,
begin,
output_begin + h_num_out,
d_num_out,
static_cast<std::int32_t>(num_items),
is_filled,
stream.get()));

// Allocate temporary storage
auto d_temp_storage = temp_allocator.allocate(temp_storage_bytes, stream);

CUCO_CUDA_TRY(cub::DeviceSelect::If(d_temp_storage,
temp_storage_bytes,
begin,
output_begin + h_num_out,
d_num_out,
static_cast<std::int32_t>(num_items),
is_filled,
stream.get()));

size_type temp_count;
CUCO_CUDA_TRY(cuco::detail::memcpy_async(
&temp_count, d_num_out, sizeof(size_type), cudaMemcpyDeviceToHost, stream));
auto const begin = cuda::make_transform_iterator(
cuda::counting_iterator{size_type{0}},
detail::open_addressing_ns::get_slot<has_payload, storage_ref_type>(this->storage_ref()));
auto const is_filled = detail::open_addressing_ns::slot_is_filled<has_payload, key_type>{
this->empty_key_sentinel(), this->erased_key_sentinel()};

std::size_t temp_storage_bytes = 0;

CUCO_CUDA_TRY(cub::DeviceSelect::If(nullptr,
temp_storage_bytes,
begin,
output_begin,
d_num_out,
this->capacity(),
is_filled,
stream.get()));

auto d_temp_storage = temp_allocator.allocate(temp_storage_bytes, stream);

CUCO_CUDA_TRY(cub::DeviceSelect::If(d_temp_storage,
temp_storage_bytes,
begin,
output_begin,
d_num_out,
this->capacity(),
is_filled,
stream.get()));
Comment thread
sleeepyjack marked this conversation as resolved.

size_type h_num_out;
CUCO_CUDA_TRY(cuco::detail::memcpy_async(
&h_num_out, d_num_out, sizeof(size_type), cudaMemcpyDeviceToHost, stream));
#if CCCL_MAJOR_VERSION > 3 || (CCCL_MAJOR_VERSION == 3 && CCCL_MINOR_VERSION >= 1)
stream.sync();
stream.sync();
#else
stream.wait();
stream.wait();
#endif
h_num_out += temp_count;
temp_allocator.deallocate(d_temp_storage, temp_storage_bytes, stream);
}

temp_allocator.deallocate(d_temp_storage, temp_storage_bytes, stream);
temp_allocator.deallocate(reinterpret_cast<char*>(d_num_out), sizeof(size_type), stream);

return output_begin + h_num_out;
Expand Down
Loading