Update MG negative sampling to return random samples distributed as specified

cpp/include/cugraph/sampling_functions.hpp

@@ -1152,7 +1152,10 @@ lookup_endpoints_from_edge_ids_and_types(
  * @param dst_biases Optional bias for randomly selecting destination vertices.  If std::nullopt
  * vertices will be selected uniformly.  In multi-GPU environment the biases should be partitioned
  * based on the vertex partitions.
- * @param num_samples Number of negative samples to generate
+ * @param num_samples Number of negative samples to generate.  In SG mode this represents the total
+ * number of samples to generate.  In MG mode, each gpu will provide the number of samples desired
+ * on that GPU.  The total number of samples in MG mode will be the aggregation of these values, the
+ * resulting samples will be randomly distributed across the ranks.
  * @param remove_duplicates If true, remove duplicate samples
  * @param remove_existing_edges If true, remove samples that are actually edges in the graph
  * @param exact_number_of_samples If true, repeat generation until we get the exact number of

cpp/src/detail/permute_range.cuh

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2024, NVIDIA CORPORATION.
+ * Copyright (c) 2024-2025, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -58,7 +58,7 @@ rmm::device_uvector<vertex_t> permute_range(raft::handle_t const& handle,
       sub_range_sizes.begin(), sub_range_sizes.end(), sub_range_sizes.begin(), global_start);
     CUGRAPH_EXPECTS(
       sub_range_sizes[comm_rank] == local_range_start,
-      "Invalid input arguments: a rage must have contiguous and non-overlapping values");
+      "Invalid input arguments: a range must have contiguous and non-overlapping values");
   }
   rmm::device_uvector<vertex_t> permuted_integers(local_range_size, handle.get_stream());
 

cpp/src/detail/permute_range_v32.cu

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2024, NVIDIA CORPORATION.
+ * Copyright (c) 2024-2025, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -16,23 +16,6 @@
 
 #include "detail/permute_range.cuh"
 
-#include <cugraph/detail/collect_comm_wrapper.hpp>
-#include <cugraph/detail/shuffle_wrappers.hpp>
-#include <cugraph/detail/utility_wrappers.hpp>
-#include <cugraph/utilities/host_scalar_comm.hpp>
-#include <cugraph/utilities/shuffle_comm.cuh>
-
-#include <raft/core/device_span.hpp>
-#include <raft/core/handle.hpp>
-#include <raft/random/rng_state.hpp>
-
-#include <rmm/device_uvector.hpp>
-#include <rmm/exec_policy.hpp>
-
-#include <thrust/iterator/constant_iterator.h>
-#include <thrust/reduce.h>
-#include <thrust/sort.h>
-
 namespace cugraph {
 
 namespace detail {

cpp/src/detail/permute_range_v64.cu

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2024, NVIDIA CORPORATION.
+ * Copyright (c) 2024-2025, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -16,23 +16,6 @@
 
 #include "detail/permute_range.cuh"
 
-#include <cugraph/detail/collect_comm_wrapper.hpp>
-#include <cugraph/detail/shuffle_wrappers.hpp>
-#include <cugraph/detail/utility_wrappers.hpp>
-#include <cugraph/utilities/host_scalar_comm.hpp>
-#include <cugraph/utilities/shuffle_comm.cuh>
-
-#include <raft/core/device_span.hpp>
-#include <raft/core/handle.hpp>
-#include <raft/random/rng_state.hpp>
-
-#include <rmm/device_uvector.hpp>
-#include <rmm/exec_policy.hpp>
-
-#include <thrust/iterator/constant_iterator.h>
-#include <thrust/reduce.h>
-#include <thrust/sort.h>
-
 namespace cugraph {
 
 namespace detail {