[INFRA] Import NVIDIA/CCCL upstream as optimization reference library

CCCL (CUDA C++ Core Libraries) provides:
- CUB: device/block/warp-level GPU primitives (reduce, scan, sort, topk)
- Thrust: high-level parallel algorithms (transform_reduce, sort, scan)
- libcudacxx: CUDA C++ standard library (atomics, barriers, memory)
- cudax: experimental features (memory resources, allocators)
- Tuning policies: per-SM hardware-specific algorithm parameters

Competition optimization vectors mapped to CCCL:
- Output TPS (83% weight): warp_reduce, block_reduce, device_topk
- Input TPS (14% weight): device_scan, block_load, prefetch
- Cache TPS (3% weight): prefix caching strategy patterns
- Memory (0.9 util): pooled/cached/buddy allocators

Source: https://github.com/NVIDIA/cccl (shallow clone, HEAD only)
License: Apache-2.0

cccl_upstream/thrust/testing/cuda/binary_search.cu

@@ -0,0 +1,24 @@
+#include <thrust/binary_search.h>
+#include <thrust/device_vector.h>
+#include <thrust/distance.h>
+#include <thrust/sequence.h>
+
+#include <cuda/std/utility>
+
+#include <unittest/unittest.h>
+
+void TestEqualRangeOnStream()
+{ // Regression test for GH issue #921 (nvbug 2173437)
+  using vector_t   = typename thrust::device_vector<int>;
+  using iterator_t = typename vector_t::iterator;
+  using result_t   = cuda::std::pair<iterator_t, iterator_t>;
+
+  vector_t input(10);
+  thrust::sequence(thrust::device, input.begin(), input.end(), 0);
+  cudaStream_t stream = nullptr;
+  result_t result     = thrust::equal_range(thrust::cuda::par.on(stream), input.begin(), input.end(), 5);
+
+  ASSERT_EQUAL(5, ::cuda::std::distance(input.begin(), result.first));
+  ASSERT_EQUAL(6, ::cuda::std::distance(input.begin(), result.second));
+}
+DECLARE_UNITTEST(TestEqualRangeOnStream);