Merge branch 'branch-25.02' into bug_debug_run

.github/workflows/pr.yaml

@@ -174,6 +174,9 @@ jobs:
     with:
       build_type: pull-request
       script: ci/test_wheel_pylibcugraph.sh
+      # CUDA 11 wheel CI is disabled until
+      # https://github.com/rapidsai/build-planning/issues/137 is resolved.
+      matrix_filter: map(select(.CUDA_VER | startswith("11") | not))
   wheel-build-cugraph:
     needs: wheel-build-pylibcugraph
     secrets: inherit
@@ -189,6 +192,9 @@ jobs:
     with:
       build_type: pull-request
       script: ci/test_wheel_cugraph.sh
+      # CUDA 11 wheel CI is disabled until
+      # https://github.com/rapidsai/build-planning/issues/137 is resolved.
+      matrix_filter: map(select(.CUDA_VER | startswith("11") | not))
   devcontainer:
     secrets: inherit
     needs: telemetry-setup

.github/workflows/test.yaml

@@ -49,6 +49,9 @@ jobs:
       date: ${{ inputs.date }}
       sha: ${{ inputs.sha }}
       script: ci/test_wheel_pylibcugraph.sh
+      # CUDA 11 wheel CI is disabled until
+      # https://github.com/rapidsai/build-planning/issues/137 is resolved.
+      matrix_filter: map(select(.CUDA_VER | startswith("11") | not))
   wheel-tests-cugraph:
     secrets: inherit
     uses: rapidsai/shared-workflows/.github/workflows/[email protected]
@@ -58,3 +61,6 @@ jobs:
       date: ${{ inputs.date }}
       sha: ${{ inputs.sha }}
       script: ci/test_wheel_cugraph.sh
+      # CUDA 11 wheel CI is disabled until
+      # https://github.com/rapidsai/build-planning/issues/137 is resolved.
+      matrix_filter: map(select(.CUDA_VER | startswith("11") | not))

ci/build_wheel.sh

@@ -40,12 +40,12 @@ EXCLUDE_ARGS=(
 # 'libraft' wheels are responsible for carrying a runtime dependency on
 # these based on RAFT's needs.
 EXCLUDE_ARGS+=(
-  --exclude "libcublas.so.12"
-  --exclude "libcublasLt.so.12"
-  --exclude "libcurand.so.10"
-  --exclude "libcusolver.so.11"
-  --exclude "libcusparse.so.12"
-  --exclude "libnvJitLink.so.12"
+  --exclude "libcublas.so.*"
+  --exclude "libcublasLt.so.*"
+  --exclude "libcurand.so.*"
+  --exclude "libcusolver.so.*"
+  --exclude "libcusparse.so.*"
+  --exclude "libnvJitLink.so.*"
 )
 
 if [[ "${package_dir}" != "python/libcugraph" ]]; then

conda/environments/all_cuda-118_arch-x86_64.yaml

@@ -35,7 +35,7 @@ dependencies:
 - networkx>=2.5.1
 - ninja
 - notebook>=0.5.0
-- numba>=0.57
+- numba>=0.59.1,<0.61.0a0
 - numpy>=1.23,<3.0a0
 - numpydoc
 - nvcc_linux-64=11.8

conda/environments/all_cuda-125_arch-x86_64.yaml

@@ -41,7 +41,7 @@ dependencies:
 - networkx>=2.5.1
 - ninja
 - notebook>=0.5.0
-- numba>=0.57
+- numba>=0.59.1,<0.61.0a0
 - numpy>=1.23,<3.0a0
 - numpydoc
 - ogb

conda/recipes/cugraph-service/meta.yaml

@@ -1,4 +1,4 @@
-# Copyright (c) 2018-2024, NVIDIA CORPORATION.
+# Copyright (c) 2018-2025, NVIDIA CORPORATION.
 
 {% set version = environ['RAPIDS_PACKAGE_VERSION'].lstrip('v') + environ.get('VERSION_SUFFIX', '') %}
 {% set minor_version = version.split('.')[0] + '.' + version.split('.')[1] %}
@@ -62,7 +62,7 @@ outputs:
         - cupy >=12.0.0
         - dask-cuda ={{ minor_version }}
         - dask-cudf ={{ minor_version }}
-        - numba >=0.57
+        - numba >=0.59.1,<0.61.0a0
         - numpy >=1.23,<3.0a0
         - python
         - rapids-dask-dependency ={{ minor_version }}

cpp/include/cugraph/detail/collect_comm_wrapper.hpp

@@ -28,11 +28,8 @@ namespace detail {
 /** @defgroup collect_comm_wrapper_cpp C++ Shuffle Wrappers
  */
 
-/** @ingroup collect_comm_wrapper_cpp
- *  @{
- */
-
 /**
+ * @ingroup collect_comm_wrapper_cpp
  * @brief Gather the span of data from all ranks and broadcast the combined data to all ranks.
  *
  * @param[in] handle RAFT handle object to encapsulate resources (e.g. CUDA stream, communicator,

cpp/include/cugraph/detail/shuffle_wrappers.hpp

@@ -28,11 +28,8 @@ namespace detail {
 /** @defgroup shuffle_wrappers_cpp C++ Shuffle Wrappers
  */
 
-/** @ingroup shuffle_wrappers_cpp
- *  @{
- */
-
 /**
+ * @ingroup shuffle_wrappers_cpp
  * @brief Shuffle external (i.e. before renumbering) vertex pairs (which can be edge end points) to
  * their local GPUs based on edge partitioning.
  *
@@ -82,6 +79,7 @@ shuffle_ext_vertex_pairs_with_values_to_local_gpu_by_edge_partitioning(
   std::optional<rmm::device_uvector<edge_time_t>>&& edge_end_times);
 
 /**
+ * @ingroup shuffle_wrappers_cpp
  * @brief Shuffle internal (i.e. renumbered) vertex pairs (which can be edge end points) to their
  * local GPUs based on edge partitioning.
  *
@@ -135,6 +133,7 @@ shuffle_int_vertex_pairs_with_values_to_local_gpu_by_edge_partitioning(
   std::vector<vertex_t> const& vertex_partition_range_lasts);
 
 /**
+ * @ingroup shuffle_wrappers_cpp
  * @brief Shuffle external (i.e. before renumbering) vertices to their local GPU based on vertex
  * partitioning.
  *
@@ -151,6 +150,7 @@ rmm::device_uvector<vertex_t> shuffle_ext_vertices_to_local_gpu_by_vertex_partit
   raft::handle_t const& handle, rmm::device_uvector<vertex_t>&& vertices);
 
 /**
+ * @ingroup shuffle_wrappers_cpp
  * @brief Shuffle external (i.e. before renumbering) vertex & value pairs to their local GPU based
  * on vertex partitioning.
  *
@@ -172,6 +172,7 @@ shuffle_ext_vertex_value_pairs_to_local_gpu_by_vertex_partitioning(
   rmm::device_uvector<value_t>&& values);
 
 /**
+ * @ingroup shuffle_wrappers_cpp
  * @brief Permute a range.
  *
  * @tparam vertex_t Type of vertex identifiers. Needs to be an integral type.
@@ -194,6 +195,7 @@ rmm::device_uvector<vertex_t> permute_range(raft::handle_t const& handle,
                                             bool do_expensive_check = false);
 
 /**
+ * @ingroup shuffle_wrappers_cpp
  * @brief Shuffle internal (i.e. renumbered) vertices to their local GPUs based on vertex
  * partitioning.
  *
@@ -214,6 +216,7 @@ rmm::device_uvector<vertex_t> shuffle_int_vertices_to_local_gpu_by_vertex_partit
   std::vector<vertex_t> const& vertex_partition_range_lasts);
 
 /**
+ * @ingroup shuffle_wrappers_cpp
  * @brief Shuffle vertices using the internal vertex key function which returns the target GPU ID.
  *
  * @tparam vertex_t Type of vertex identifiers. Needs to be an integral type.
@@ -236,6 +239,7 @@ shuffle_int_vertex_value_pairs_to_local_gpu_by_vertex_partitioning(
   std::vector<vertex_t> const& vertex_partition_range_lasts);
 
 /**
+ * @ingroup shuffle_wrappers_cpp
  * @brief Groupby and count edgelist using the key function which returns the target local partition
  * ID for an edge.
  *
@@ -282,6 +286,7 @@ rmm::device_uvector<size_t> groupby_and_count_edgelist_by_local_partition_id(
   bool groupby_and_count_local_partition_by_minor = false);
 
 /**
+ * @ingroup shuffle_wrappers_cpp
  * @brief Collect vertex values (represented as k/v pairs across cluster) and return
  *        local value arrays on the GPU responsible for each vertex.
  *
@@ -317,7 +322,3 @@ rmm::device_uvector<value_t> collect_local_vertex_values_from_ext_vertex_value_p
 
 }  // namespace detail
 }  // namespace cugraph
-
-/**
- * @}
- */

cpp/include/cugraph/detail/utility_wrappers.hpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021-2024, NVIDIA CORPORATION.
+ * Copyright (c) 2021-2025, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -26,7 +26,11 @@
 namespace cugraph {
 namespace detail {
 
+/** @defgroup utility_wrappers_cpp C++ Utility Wrappers
+ */
+
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief    Fill a buffer with uniformly distributed random values
  *
  * Fills a buffer with uniformly distributed random values between
@@ -51,6 +55,7 @@ void uniform_random_fill(rmm::cuda_stream_view const& stream_view,
                          raft::random::RngState& rng_state);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief    Fill a buffer with a constant value
  *
  * @tparam      value_t      type of the value to operate on
@@ -66,6 +71,7 @@ template <typename value_t>
 void scalar_fill(raft::handle_t const& handle, value_t* d_value, size_t size, value_t value);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief    Sort a device span
  *
  * @tparam      value_t      type of the value to operate on. Must be either int32_t or int64_t.
@@ -79,6 +85,7 @@ template <typename value_t>
 void sort_ints(raft::handle_t const& handle, raft::device_span<value_t> values);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief    Keep unique element from a device span
  *
  * @tparam      value_t      type of the value to operate on. Must be either int32_t or int64_t.
@@ -93,6 +100,7 @@ template <typename value_t>
 size_t unique_ints(raft::handle_t const& handle, raft::device_span<value_t> values);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief    Increment the values of a device span by a constant value
  *
  * @tparam      value_t      type of the value to operate on. Must be either int32_t or int64_t.
@@ -108,6 +116,7 @@ void transform_increment_ints(raft::device_span<value_t> values,
                               rmm::cuda_stream_view const& stream_view);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief    Fill a buffer with a sequence of values
  *
  * Fills the buffer with the sequence:
@@ -130,6 +139,7 @@ void sequence_fill(rmm::cuda_stream_view const& stream_view,
                    value_t start_value);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief    Fill a buffer with a sequence of values with the input stride
  *
  * Fills the buffer with the sequence with the input stride:
@@ -152,6 +162,7 @@ void stride_fill(rmm::cuda_stream_view const& stream_view,
                  value_t stride);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief    Compute the maximum vertex id of an edge list
  *
  * max(d_edgelist_srcs.max(), d_edgelist_dsts.max())
@@ -172,6 +183,7 @@ vertex_t compute_maximum_vertex_id(rmm::cuda_stream_view const& stream_view,
                                    size_t num_edges);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief    Compute the maximum vertex id of an edge list
  *
  * max(d_edgelist_srcs.max(), d_edgelist_dsts.max())
@@ -194,6 +206,7 @@ vertex_t compute_maximum_vertex_id(rmm::cuda_stream_view const& stream_view,
 }
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief Filter zero degree vertices from this frontier
  *
  * @tparam vertex_t vertex type
@@ -211,6 +224,7 @@ std::tuple<rmm::device_uvector<vertex_t>, rmm::device_uvector<edge_t>> filter_de
   rmm::device_uvector<edge_t>&& d_out_degs);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief Check if device span is sorted
  *
  * @tparam data_t type of data in span
@@ -223,6 +237,7 @@ template <typename data_t>
 bool is_sorted(raft::handle_t const& handle, raft::device_span<data_t> span);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief Check if two device spans are equal.  Returns true if every element in the spans are
  * equal.
  *
@@ -239,6 +254,7 @@ bool is_equal(raft::handle_t const& handle,
               raft::device_span<data_t> span2);
 
 /**
+ * @ingroup utility_wrappers_cpp
  * @brief Count the number of times a value appears in a span
  *
  * @tparam data_t type of data in span