[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/functional_placeholders_miscellaneous.cu

@@ -0,0 +1,109 @@
+#include <thrust/functional.h>
+#include <thrust/transform.h>
+
+#include <unittest/unittest.h>
+
+template <typename T>
+struct saxpy_reference
+{
+  _CCCL_HOST_DEVICE saxpy_reference(const T& aa)
+      : a(aa)
+  {}
+
+  _CCCL_HOST_DEVICE T operator()(const T& x, const T& y) const
+  {
+    return a * x + y;
+  }
+
+  T a;
+};
+
+template <typename Vector>
+struct TestFunctionalPlaceholdersValue
+{
+  void operator()(const size_t)
+  {
+    const size_t n = 10000;
+    using T        = typename Vector::value_type;
+
+    T a(13);
+
+    Vector x = unittest::random_integers<T>(n);
+    Vector y = unittest::random_integers<T>(n);
+    Vector result(n), reference(n);
+
+    thrust::transform(x.begin(), x.end(), y.begin(), reference.begin(), saxpy_reference<T>(a));
+
+    using namespace thrust::placeholders;
+    thrust::transform(x.begin(), x.end(), y.begin(), result.begin(), a * _1 + _2);
+
+    ASSERT_ALMOST_EQUAL(reference, result);
+  }
+};
+VectorUnitTest<TestFunctionalPlaceholdersValue, ThirtyTwoBitTypes, thrust::device_vector, thrust::device_allocator>
+  TestFunctionalPlaceholdersValueDevice;
+VectorUnitTest<TestFunctionalPlaceholdersValue, ThirtyTwoBitTypes, thrust::host_vector, std::allocator>
+  TestFunctionalPlaceholdersValueHost;
+
+template <typename Vector>
+struct TestFunctionalPlaceholdersTransformIterator
+{
+  void operator()(const size_t)
+  {
+    const size_t n = 10000;
+    using T        = typename Vector::value_type;
+
+    T a(13);
+
+    Vector x = unittest::random_integers<T>(n);
+    Vector y = unittest::random_integers<T>(n);
+    Vector result(n), reference(n);
+
+    thrust::transform(x.begin(), x.end(), y.begin(), reference.begin(), saxpy_reference<T>(a));
+
+    using namespace thrust::placeholders;
+    thrust::transform(
+      thrust::make_transform_iterator(x.begin(), a * _1),
+      thrust::make_transform_iterator(x.end(), a * _1),
+      y.begin(),
+      result.begin(),
+      _1 + _2);
+
+    ASSERT_ALMOST_EQUAL(reference, result);
+  }
+};
+VectorUnitTest<TestFunctionalPlaceholdersTransformIterator,
+               ThirtyTwoBitTypes,
+               thrust::device_vector,
+               thrust::device_allocator>
+  TestFunctionalPlaceholdersTransformIteratorInstanceDevice;
+VectorUnitTest<TestFunctionalPlaceholdersTransformIterator, ThirtyTwoBitTypes, thrust::host_vector, std::allocator>
+  TestFunctionalPlaceholdersTransformIteratorInstanceHost;
+
+void TestFunctionalPlaceholdersArgumentValueCategories()
+{
+  using namespace thrust::placeholders;
+  auto expr = _1 * _1 + _2 * _2;
+  int a     = 2;
+  int b     = 3;
+  ASSERT_EQUAL(expr(2, 3), 13); // pass pr-value
+  ASSERT_EQUAL(expr(a, b), 13); // pass l-value
+  ASSERT_EQUAL(expr(::cuda::std::move(a), ::cuda::std::move(b)), 13); // pass x-value
+}
+DECLARE_UNITTEST(TestFunctionalPlaceholdersArgumentValueCategories);
+
+void TestFunctionalPlaceholdersSemiRegular()
+{
+  using namespace thrust::placeholders;
+  using Expr = decltype(_1 * _1 + _2 * _2);
+  Expr expr; // default-constructible
+  ASSERT_EQUAL(expr(2, 3), 13);
+  Expr expr2 = expr; // copy-constructible
+  ASSERT_EQUAL(expr2(2, 3), 13);
+  Expr expr3;
+  expr3 = expr; // copy-assignable
+  ASSERT_EQUAL(expr3(2, 3), 13);
+
+  static_assert(::cuda::std::semiregular<Expr>);
+}
+DECLARE_UNITTEST(TestFunctionalPlaceholdersSemiRegular);