Skip to content

Fix failing CAGRA merge google tests for 25.06 #974

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
raydouglass merged 7 commits into from
Jun 4, 2025
Merged

Fix failing CAGRA merge google tests for 25.06 #974

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
6339868
Fix the launch bounds for nn-descent kernel for 1210
achirkin Jun 2, 2025
bb13eb3
Invert the CUDA_ARCH logic
achirkin Jun 2, 2025
3f7c9c4
Fix ==
bdice Jun 2, 2025
5ccbfbe
Fix style
bdice Jun 2, 2025
6c7e325
Update cpp/src/neighbors/detail/nn_descent.cuh
achirkin Jun 3, 2025
78a1651
Merge branch 'branch-25.06' into achirkin-backport-nn-descent-launch-…
achirkin Jun 3, 2025
fe27ae6
revert to host in one memory
rhdong Jun 4, 2025
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
9 changes: 5 additions & 4 deletions cpp/src/neighbors/detail/nn_descent.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -491,11 +491,12 @@ __device__ __forceinline__ void remove_duplicates(
template <typename Index_t, typename ID_t = InternalID_t<Index_t>>
RAFT_KERNEL
#ifdef __CUDA_ARCH__
#if (__CUDA_ARCH__) == 750 || ((__CUDA_ARCH__) >= 860 && (__CUDA_ARCH__) <= 890) || \
(__CUDA_ARCH__) == 1200
__launch_bounds__(BLOCK_SIZE)
#else
// Use minBlocksPerMultiprocessor = 4 on specific arches
#if (__CUDA_ARCH__) == 700 || (__CUDA_ARCH__) == 800 || (__CUDA_ARCH__) == 900 || \
(__CUDA_ARCH__) == 1000
__launch_bounds__(BLOCK_SIZE, 4)
#else
__launch_bounds__(BLOCK_SIZE)
#endif
#endif
local_join_kernel(const Index_t* graph_new,
Expand Down
23 changes: 8 additions & 15 deletions cpp/tests/neighbors/ann_cagra.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1035,29 +1035,22 @@ class AnnCagraIndexMergeTest : public ::testing::TestWithParam<AnnCagraInputs> {
auto database1_view = raft::make_device_matrix_view<const DataT, int64_t>(
(const DataT*)database.data() + database0_view.size(), database1_size, ps.dim);

cagra::index<DataT, IdxT> index0(handle_);
cagra::index<DataT, IdxT> index1(handle_);
cagra::index<DataT, IdxT> index0(handle_, index_params.metric);
cagra::index<DataT, IdxT> index1(handle_, index_params.metric);
std::optional<raft::host_matrix<DataT, int64_t>> database_host{std::nullopt};
if (ps.host_dataset) {
database_host = raft::make_host_matrix<DataT, int64_t>(handle_, ps.n_rows, ps.dim);
raft::copy(database_host->data_handle(), database.data(), database.size(), stream_);
{
std::optional<raft::host_matrix<DataT, int64_t>> database_host{std::nullopt};
database_host = raft::make_host_matrix<DataT, int64_t>(database0_size, ps.dim);
raft::copy(database_host->data_handle(),
database0_view.data_handle(),
database0_view.size(),
stream_);
auto database_host_view = raft::make_host_matrix_view<const DataT, int64_t>(
(const DataT*)database_host->data_handle(), database0_size, ps.dim);
index0 = cagra::build(handle_, index_params, database_host_view);
}
{
std::optional<raft::host_matrix<DataT, int64_t>> database_host{std::nullopt};
database_host = raft::make_host_matrix<DataT, int64_t>(database1_size, ps.dim);
raft::copy(database_host->data_handle(),
database1_view.data_handle(),
database1_view.size(),
stream_);
auto database_host_view = raft::make_host_matrix_view<const DataT, int64_t>(
(const DataT*)database_host->data_handle(), database1_size, ps.dim);
(const DataT*)database_host->data_handle() + database0_size * ps.dim,
database1_size,
ps.dim);
index1 = cagra::build(handle_, index_params, database_host_view);
}
} else {
Expand Down