Add random subsampling for IVF methods

cpp/include/raft/matrix/detail/gather.cuh

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
+ * Copyright (c) 2022-2024, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -17,7 +17,15 @@
 #pragma once
 
 #include <functional>
+#include <raft/common/nvtx.hpp>
+#include <raft/core/device_mdarray.hpp>
+#include <raft/core/device_mdspan.hpp>
+#include <raft/core/host_mdarray.hpp>
+#include <raft/core/host_mdspan.hpp>
 #include <raft/core/operators.hpp>
+#include <raft/core/pinned_mdarray.hpp>
+#include <raft/core/pinned_mdspan.hpp>
+#include <raft/util/cuda_dev_essentials.cuh>
 #include <raft/util/cudart_utils.hpp>
 
 namespace raft {
@@ -335,6 +343,70 @@ void gather_if(const InputIteratorT in,
   gatherImpl(in, D, N, map, stencil, map_length, out, pred_op, transform_op, stream);
 }
 
+template <typename T, typename IdxT = int64_t>
+void gather_buff(host_matrix_view<const T, IdxT> dataset,
+                 host_vector_view<const IdxT, IdxT> indices,
+                 IdxT offset,
+                 pinned_matrix_view<T, IdxT> buff)
+{
+  raft::common::nvtx::range<common::nvtx::domain::raft> fun_scope("gather_host_buff");
+  IdxT batch_size = std::min<IdxT>(buff.extent(0), indices.extent(0) - offset);
+
+#pragma omp for
+  for (IdxT i = 0; i < batch_size; i++) {
+    IdxT in_idx = indices(offset + i);
+    for (IdxT k = 0; k < buff.extent(1); k++) {
+      buff(i, k) = dataset(in_idx, k);
+    }
+  }
+}
+
+template <typename T, typename IdxT>
+void gather(raft::resources const& res,
+            host_matrix_view<const T, IdxT> dataset,
+            device_vector_view<const IdxT, IdxT> indices,
+            raft::device_matrix_view<T, IdxT> output)
+{
+  raft::common::nvtx::range<common::nvtx::domain::raft> fun_scope("gather");
+  IdxT n_dim        = output.extent(1);
+  IdxT n_train      = output.extent(0);
+  auto indices_host = raft::make_host_vector<IdxT, IdxT>(n_train);
+  raft::copy(
+    indices_host.data_handle(), indices.data_handle(), n_train, resource::get_cuda_stream(res));
+  resource::sync_stream(res);
+
+  const size_t max_batch_size = 32768;
+  // Gather the vector on the host in tmp buffers. We use two buffers to overlap H2D sync
+  // and gathering the data.
+  raft::common::nvtx::push_range("gather::alloc_buffers");
+  auto out_tmp1 = raft::make_pinned_matrix<T, IdxT>(res, max_batch_size, n_dim);
+  auto out_tmp2 = raft::make_pinned_matrix<T, IdxT>(res, max_batch_size, n_dim);
+  auto view1    = out_tmp1.view();
+  auto view2    = out_tmp2.view();
+  raft::common::nvtx::pop_range();
+
+  gather_buff(dataset, make_const_mdspan(indices_host.view()), (IdxT)0, view1);
+#pragma omp parallel
+  for (IdxT device_offset = 0; device_offset < n_train; device_offset += max_batch_size) {
+    IdxT batch_size = std::min<IdxT>(max_batch_size, n_train - device_offset);
+#pragma omp master
+    raft::copy(output.data_handle() + device_offset * n_dim,
+               view1.data_handle(),
+               batch_size * n_dim,
+               resource::get_cuda_stream(res));
+    // Start gathering the next batch on the host.
+    IdxT host_offset = device_offset + batch_size;
+    batch_size       = std::min<IdxT>(max_batch_size, n_train - host_offset);
+    if (batch_size > 0) {
+      gather_buff(dataset, make_const_mdspan(indices_host.view()), host_offset, view2);
+    }
+#pragma omp master
+    resource::sync_stream(res);
+#pragma omp barrier
+    std::swap(view1, view2);
+  }
+}
+
 }  // namespace detail
 }  // namespace matrix
 }  // namespace raft

cpp/include/raft/neighbors/detail/ivf_flat_build.cuh

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
+ * Copyright (c) 2022-2024, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -361,28 +361,23 @@ inline auto build(raft::resources const& handle,
 
   // Train the kmeans clustering
   {
+    int random_seed     = 137;
     auto trainset_ratio = std::max<size_t>(
       1, n_rows / std::max<size_t>(params.kmeans_trainset_fraction * n_rows, index.n_lists()));
     auto n_rows_train = n_rows / trainset_ratio;
-    rmm::device_uvector<T> trainset(n_rows_train * index.dim(), stream);
-    // TODO: a proper sampling
-    RAFT_CUDA_TRY(cudaMemcpy2DAsync(trainset.data(),
-                                    sizeof(T) * index.dim(),
-                                    dataset,
-                                    sizeof(T) * index.dim() * trainset_ratio,
-                                    sizeof(T) * index.dim(),
-                                    n_rows_train,
-                                    cudaMemcpyDefault,
-                                    stream));
-    auto trainset_const_view =
-      raft::make_device_matrix_view<const T, IdxT>(trainset.data(), n_rows_train, index.dim());
+    auto trainset     = make_device_matrix<T, IdxT>(handle, n_rows_train, index.dim());
+    raft::spatial::knn::detail::utils::subsample(
+      handle, dataset, n_rows, trainset.view(), random_seed);
     auto centers_view = raft::make_device_matrix_view<float, IdxT>(
       index.centers().data_handle(), index.n_lists(), index.dim());
     raft::cluster::kmeans_balanced_params kmeans_params;
     kmeans_params.n_iters = params.kmeans_n_iters;
     kmeans_params.metric  = index.metric();
-    raft::cluster::kmeans_balanced::fit(
-      handle, kmeans_params, trainset_const_view, centers_view, utils::mapping<float>{});
+    raft::cluster::kmeans_balanced::fit(handle,
+                                        kmeans_params,
+                                        make_const_mdspan(trainset.view()),
+                                        centers_view,
+                                        utils::mapping<float>{});
   }
 
   // add the data if necessary

cpp/include/raft/neighbors/detail/ivf_pq_build.cuh

@@ -25,9 +25,11 @@
 
 #include <raft/cluster/kmeans_balanced.cuh>
 #include <raft/core/device_mdarray.hpp>
+#include <raft/core/device_resources.hpp>
 #include <raft/core/logger.hpp>
 #include <raft/core/nvtx.hpp>
 #include <raft/core/operators.hpp>
+#include <raft/core/resource/device_memory_resource.hpp>
 #include <raft/core/resources.hpp>
 #include <raft/distance/distance_types.hpp>
 #include <raft/linalg/add.cuh>
@@ -46,7 +48,6 @@
 #include <raft/util/pow2_utils.cuh>
 #include <raft/util/vectorized.cuh>
 
-#include <raft/core/resource/device_memory_resource.hpp>
 #include <rmm/cuda_stream_view.hpp>
 #include <rmm/device_uvector.hpp>
 #include <rmm/mr/device/managed_memory_resource.hpp>
@@ -1754,76 +1755,47 @@ auto build(raft::resources const& handle,
   utils::memzero(index.inds_ptrs().data_handle(), index.inds_ptrs().size(), stream);
 
   {
+    int random_seed     = 137;
     auto trainset_ratio = std::max<size_t>(
       1,
       size_t(n_rows) / std::max<size_t>(params.kmeans_trainset_fraction * n_rows, index.n_lists()));
     size_t n_rows_train = n_rows / trainset_ratio;
 
-    auto* device_memory = resource::get_workspace_resource(handle);
-    rmm::mr::managed_memory_resource managed_memory_upstream;
+    auto* device_mr = resource::get_workspace_resource(handle);
+    rmm::mr::managed_memory_resource managed_mr;
 
     // Besides just sampling, we transform the input dataset into floats to make it easier
     // to use gemm operations from cublas.
-    rmm::device_uvector<float> trainset(n_rows_train * index.dim(), stream, device_memory);
-    // TODO: a proper sampling
+    auto trainset =
+      make_device_mdarray<float>(handle, device_mr, make_extents<IdxT>(n_rows_train, dim));
+
     if constexpr (std::is_same_v<T, float>) {
-      RAFT_CUDA_TRY(cudaMemcpy2DAsync(trainset.data(),
-                                      sizeof(T) * index.dim(),
-                                      dataset,
-                                      sizeof(T) * index.dim() * trainset_ratio,
-                                      sizeof(T) * index.dim(),
-                                      n_rows_train,
-                                      cudaMemcpyDefault,
-                                      stream));
+      raft::spatial::knn::detail::utils::subsample(
+        handle, dataset, n_rows, trainset.view(), random_seed);
     } else {
-      size_t dim = index.dim();
-      cudaPointerAttributes dataset_attr;
-      RAFT_CUDA_TRY(cudaPointerGetAttributes(&dataset_attr, dataset));
-      if (dataset_attr.devicePointer != nullptr) {
-        // data is available on device: just run the kernel to copy and map the data
-        auto p = reinterpret_cast<T*>(dataset_attr.devicePointer);
-        auto trainset_view =
-          raft::make_device_vector_view<float, IdxT>(trainset.data(), dim * n_rows_train);
-        linalg::map_offset(handle, trainset_view, [p, trainset_ratio, dim] __device__(size_t i) {
-          auto col = i % dim;
-          return utils::mapping<float>{}(p[(i - col) * size_t(trainset_ratio) + col]);
-        });
-      } else {
-        // data is not available: first copy, then map inplace
-        auto trainset_tmp = reinterpret_cast<T*>(reinterpret_cast<uint8_t*>(trainset.data()) +
-                                                 (sizeof(float) - sizeof(T)) * index.dim());
-        // We copy the data in strides, one row at a time, and place the smaller rows of type T
-        // at the end of float rows.
-        RAFT_CUDA_TRY(cudaMemcpy2DAsync(trainset_tmp,
-                                        sizeof(float) * index.dim(),
-                                        dataset,
-                                        sizeof(T) * index.dim() * trainset_ratio,
-                                        sizeof(T) * index.dim(),
-                                        n_rows_train,
-                                        cudaMemcpyDefault,
-                                        stream));
-        // Transform the input `{T -> float}`, one row per warp.
-        // The threads in each warp copy the data synchronously; this and the layout of the data
-        // (content is aligned to the end of the rows) together allow doing the transform in-place.
-        copy_warped(trainset.data(),
-                    index.dim(),
-                    trainset_tmp,
-                    index.dim() * sizeof(float) / sizeof(T),
-                    index.dim(),
-                    n_rows_train,
-                    stream);
-      }
+      // TODO(tfeher): Enable codebook generation with any type T, and then remove
+      // trainset tmp.
+      auto trainset_tmp =
+        make_device_mdarray<T>(handle, &managed_mr, make_extents<IdxT>(n_rows_train, dim));
+      raft::spatial::knn::detail::utils::subsample(
+        handle, dataset, n_rows, trainset_tmp.view(), random_seed);
+      cudaDeviceSynchronize();
+      RAFT_LOG_INFO("Subsampling done, converting to float");
+      raft::linalg::unaryOp(trainset.data_handle(),
+                            trainset_tmp.data_handle(),
+                            trainset.size(),
+                            utils::mapping<float>{},  // raft::cast_op<float>(),
+                            raft::resource::get_cuda_stream(handle));
     }
 
     // NB: here cluster_centers is used as if it is [n_clusters, data_dim] not [n_clusters,
     // dim_ext]!
     rmm::device_uvector<float> cluster_centers_buf(
-      index.n_lists() * index.dim(), stream, device_memory);
+      index.n_lists() * index.dim(), stream, device_mr);
     auto cluster_centers = cluster_centers_buf.data();
 
     // Train balanced hierarchical kmeans clustering
-    auto trainset_const_view =
-      raft::make_device_matrix_view<const float, IdxT>(trainset.data(), n_rows_train, index.dim());
+    auto trainset_const_view = raft::make_const_mdspan(trainset.view());
     auto centers_view =
       raft::make_device_matrix_view<float, IdxT>(cluster_centers, index.n_lists(), index.dim());
     raft::cluster::kmeans_balanced_params kmeans_params;
@@ -1833,7 +1805,7 @@ auto build(raft::resources const& handle,
       handle, kmeans_params, trainset_const_view, centers_view, utils::mapping<float>{});
 
     // Trainset labels are needed for training PQ codebooks
-    rmm::device_uvector<uint32_t> labels(n_rows_train, stream, device_memory);
+    rmm::device_uvector<uint32_t> labels(n_rows_train, stream, device_mr);
     auto centers_const_view = raft::make_device_matrix_view<const float, IdxT>(
       cluster_centers, index.n_lists(), index.dim());
     auto labels_view = raft::make_device_vector_view<uint32_t, IdxT>(labels.data(), n_rows_train);
@@ -1859,19 +1831,19 @@ auto build(raft::resources const& handle,
         train_per_subset(handle,
                          index,
                          n_rows_train,
-                         trainset.data(),
+                         trainset.data_handle(),
                          labels.data(),
                          params.kmeans_n_iters,
-                         &managed_memory_upstream);
+                         &managed_mr);
         break;
       case codebook_gen::PER_CLUSTER:
         train_per_cluster(handle,
                           index,
                           n_rows_train,
-                          trainset.data(),
+                          trainset.data_handle(),
                           labels.data(),
                           params.kmeans_n_iters,
-                          &managed_memory_upstream);
+                          &managed_mr);
         break;
       default: RAFT_FAIL("Unreachable code");
     }

cpp/include/raft/spatial/knn/detail/ann_utils.cuh

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
+ * Copyright (c) 2022-2024, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -16,8 +16,16 @@
 
 #pragma once
 
+#include <raft/common/nvtx.hpp>
+#include <raft/core/device_mdarray.hpp>
+#include <raft/core/device_mdspan.hpp>
+#include <raft/core/device_resources.hpp>
+#include <raft/core/host_mdarray.hpp>
+
 #include <raft/core/logger.hpp>
 #include <raft/distance/distance_types.hpp>
+#include <raft/matrix/gather.cuh>
+#include <raft/random/sample_without_replacement.cuh>
 #include <raft/util/cuda_utils.cuh>
 #include <raft/util/cudart_utils.hpp>
 #include <raft/util/integer_utils.hpp>
@@ -30,6 +38,10 @@
 #include <optional>
 
 #include <cuda_fp16.hpp>
+#include <thrust/copy.h>
+#include <thrust/device_ptr.h>
+#include <thrust/iterator/counting_iterator.h>
+#include <thrust/sort.h>
 
 namespace raft::spatial::knn::detail::utils {
 
@@ -573,4 +585,52 @@ struct batch_load_iterator {
   size_type cur_pos_;
 };
 
+template <typename IdxT>
+auto get_subsample_indices(raft::resources const& res, IdxT n_samples, IdxT n_subsamples, int seed)
+  -> raft::device_vector<IdxT, IdxT>
+{
+  RAFT_EXPECTS(n_subsamples <= n_samples, "Cannot have more training samples than dataset vectors");
+
+  auto data_indices = raft::make_device_vector<IdxT, IdxT>(res, n_samples);
+  thrust::counting_iterator<IdxT> first(0);
+  thrust::device_ptr<IdxT> ptr(data_indices.data_handle());
+  thrust::copy(raft::resource::get_thrust_policy(res), first, first + n_samples, ptr);
+  raft::random::RngState rng(seed);
+  auto train_indices = raft::make_device_vector<IdxT, IdxT>(res, n_subsamples);
+  raft::random::sample_without_replacement(res,
+                                           rng,
+                                           raft::make_const_mdspan(data_indices.view()),
+                                           std::nullopt,
+                                           train_indices.view(),
+                                           std::nullopt);
+  return train_indices;
+}
+
+/** Subsample the dataset to create a training set*/
+template <typename T, typename IdxT = int64_t>
+void subsample(raft::resources const& res,
+               const T* input,
+               IdxT n_samples,
+               raft::device_matrix_view<T, IdxT> output,
+               int seed)
+{
+  IdxT n_dim   = output.extent(1);
+  IdxT n_train = output.extent(0);
+
+  raft::device_vector<IdxT, IdxT> train_indices =
+    get_subsample_indices<IdxT>(res, n_samples, n_train, seed);
+
+  cudaPointerAttributes attr;
+  RAFT_CUDA_TRY(cudaPointerGetAttributes(&attr, input));
+  T* ptr = reinterpret_cast<T*>(attr.devicePointer);
+  if (ptr != nullptr) {
+    raft::matrix::gather(res,
+                         raft::make_device_matrix_view<const T, IdxT>(ptr, n_samples, n_dim),
+                         raft::make_const_mdspan(train_indices.view()),
+                         output);
+  } else {
+    auto dataset = raft::make_host_matrix_view<const T, IdxT>(input, n_samples, n_dim);
+    raft::matrix::detail::gather(res, dataset, make_const_mdspan(train_indices.view()), output);
+  }
+}
 }  // namespace raft::spatial::knn::detail::utils

cpp/test/neighbors/ann_ivf_pq.cuh

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
+ * Copyright (c) 2022-2024, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -68,7 +68,7 @@ struct ivf_pq_inputs {
   ivf_pq_inputs()
   {
     index_params.n_lists                  = max(32u, min(1024u, num_db_vecs / 128u));
-    index_params.kmeans_trainset_fraction = 1.0;
+    index_params.kmeans_trainset_fraction = 0.95;
   }
 };