Vamana build optimization set and fp16 support

c/src/neighbors/vamana.cpp

@@ -1,9 +1,10 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  */
 
 #include <cstdint>
+#include <cuda_fp16.h>
 #include <dlpack/dlpack.h>
 
 #include <raft/core/error.hpp>
@@ -82,6 +83,8 @@ extern "C" cuvsError_t cuvsVamanaIndexDestroy(cuvsVamanaIndex_t index_c_ptr)
     if (index.addr != 0) {
       if (index.dtype.code == kDLFloat && index.dtype.bits == 32) {
         delete reinterpret_cast<cuvs::neighbors::vamana::index<float, uint32_t>*>(index.addr);
+      } else if (index.dtype.code == kDLFloat && index.dtype.bits == 16) {
+        delete reinterpret_cast<cuvs::neighbors::vamana::index<half, uint32_t>*>(index.addr);
       } else if (index.dtype.code == kDLInt && index.dtype.bits == 8) {
         delete reinterpret_cast<cuvs::neighbors::vamana::index<int8_t, uint32_t>*>(index.addr);
       } else if (index.dtype.code == kDLUInt && index.dtype.bits == 8) {
@@ -100,6 +103,10 @@ extern "C" cuvsError_t cuvsVamanaIndexGetDims(cuvsVamanaIndex_t index, int* dim)
       auto index_ptr =
         reinterpret_cast<cuvs::neighbors::vamana::index<float, uint32_t>*>(index->addr);
       *dim = index_ptr->dim();
+    } else if (index->dtype.code == kDLFloat && index->dtype.bits == 16) {
+      auto index_ptr =
+        reinterpret_cast<cuvs::neighbors::vamana::index<half, uint32_t>*>(index->addr);
+      *dim = index_ptr->dim();
     } else if (index->dtype.code == kDLInt && index->dtype.bits == 8) {
       auto index_ptr =
         reinterpret_cast<cuvs::neighbors::vamana::index<int8_t, uint32_t>*>(index->addr);
@@ -123,6 +130,8 @@ extern "C" cuvsError_t cuvsVamanaBuild(cuvsResources_t res,
 
     if (dataset.dtype.code == kDLFloat && dataset.dtype.bits == 32) {
       index->addr = reinterpret_cast<uintptr_t>(_build<float>(res, params, dataset_tensor));
+    } else if (dataset.dtype.code == kDLFloat && dataset.dtype.bits == 16) {
+      index->addr = reinterpret_cast<uintptr_t>(_build<half>(res, params, dataset_tensor));
     } else if (dataset.dtype.code == kDLInt && dataset.dtype.bits == 8) {
       index->addr = reinterpret_cast<uintptr_t>(_build<int8_t>(res, params, dataset_tensor));
     } else if (dataset.dtype.code == kDLUInt && dataset.dtype.bits == 8) {
@@ -143,6 +152,8 @@ extern "C" cuvsError_t cuvsVamanaSerialize(cuvsResources_t res,
   return cuvs::core::translate_exceptions([=] {
     if (index->dtype.code == kDLFloat && index->dtype.bits == 32) {
       _serialize<float>(res, filename, index, include_dataset);
+    } else if (index->dtype.code == kDLFloat && index->dtype.bits == 16) {
+      _serialize<half>(res, filename, index, include_dataset);
     } else if (index->dtype.code == kDLInt && index->dtype.bits == 8) {
       _serialize<int8_t>(res, filename, index, include_dataset);
     } else if (index->dtype.code == kDLUInt && index->dtype.bits == 8) {

cpp/CMakeLists.txt

@@ -1412,10 +1412,12 @@ if(NOT BUILD_CPU_ONLY)
     src/neighbors/tiered_index.cu
     src/neighbors/sparse_brute_force.cu
     src/neighbors/vamana_build_float.cu
+    src/neighbors/vamana_build_half.cu
     src/neighbors/vamana_build_uint8.cu
     src/neighbors/vamana_build_int8.cu
     src/neighbors/vamana_codebooks_float.cu
     src/neighbors/vamana_serialize_float.cu
+    src/neighbors/vamana_serialize_half.cu
     src/neighbors/vamana_serialize_uint8.cu
     src/neighbors/vamana_serialize_int8.cu
     src/preprocessing/quantize/scalar.cu

cpp/include/cuvs/neighbors/vamana.hpp

@@ -1,11 +1,12 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  */
 
 #pragma once
 
 #include "common.hpp"
+#include <cuda_fp16.h>
 #include <cuvs/distance/distance.hpp>
 #include <cuvs/neighbors/common.hpp>
 #include <raft/core/device_mdspan.hpp>
@@ -344,6 +345,16 @@ auto build(raft::resources const& res,
            raft::host_matrix_view<const float, int64_t, raft::row_major> dataset)
   -> cuvs::neighbors::vamana::index<float, uint32_t>;
 
+auto build(raft::resources const& res,
+           const cuvs::neighbors::vamana::index_params& params,
+           raft::device_matrix_view<const half, int64_t, raft::row_major> dataset)
+  -> cuvs::neighbors::vamana::index<half, uint32_t>;
+
+auto build(raft::resources const& res,
+           const cuvs::neighbors::vamana::index_params& params,
+           raft::host_matrix_view<const half, int64_t, raft::row_major> dataset)
+  -> cuvs::neighbors::vamana::index<half, uint32_t>;
+
 /**
  * @brief Build the index from the dataset for efficient DiskANN search.
  *
@@ -520,6 +531,12 @@ void serialize(raft::resources const& handle,
                bool include_dataset = true,
                bool sector_aligned  = false);
 
+void serialize(raft::resources const& handle,
+               const std::string& file_prefix,
+               const cuvs::neighbors::vamana::index<half, uint32_t>& index,
+               bool include_dataset = true,
+               bool sector_aligned  = false);
+
 /**
  * Save the index to file.
  *