Prevent nested parallelism in HNSW bench

- Setting the gbench number of threads and the HNSWlib config number of threads can lead to nested parallelism. Force either throughput mode using multiple gbench threads or latency mode using batch paralleism.
- Added a check in `search` method to handle single query batch size efficiently. There is a significant overhead in going throught he thread pool.

Author: julianmi

cpp/bench/ann/src/hnswlib/hnswlib_wrapper.h

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2023-2024, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2023-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 #pragma once
@@ -169,8 +169,12 @@ void hnsw_lib<T>::set_search_param(const search_param_base& param_, const void*
   auto param     = dynamic_cast<const search_param&>(param_);
   appr_alg_->ef_ = param.ef;
   num_threads_   = param.num_threads;
-  // bench_mode_ = param.metric_objective;
-  bench_mode_ = Mode::kLatency;  // TODO(achirkin): pass the benchmark mode in the algo parameters
+  if (cuvs::bench::benchmark_n_threads > 1) {
+    bench_mode_  = Mode::kThroughput;
+    num_threads_ = 1;  // Prevent nested parallelism (gbench threads + batch threads).
+  } else {
+    bench_mode_ = Mode::kLatency;
+  }
 
   // Create a pool if multiple query threads have been set and the pool hasn't been created already
   bool create_pool = (bench_mode_ == Mode::kLatency && num_threads_ > 1 && !thread_pool_);
@@ -181,6 +185,11 @@ template <typename T>
 void hnsw_lib<T>::search(
   const T* query, int batch_size, int k, algo_base::index_type* indices, float* distances) const
 {
+  if (batch_size == 1) {
+    get_search_knn_results(query, k, indices, distances);
+    return;
+  }
+
   auto f = [&](int i) {
     // hnsw can only handle a single vector at a time.
     get_search_knn_results(query + i * dim_, k, indices + i * k, distances + i * k);