[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/tabulate.cu

@@ -0,0 +1,94 @@
+#include <thrust/functional.h>
+#include <thrust/iterator/discard_iterator.h>
+#include <thrust/iterator/retag.h>
+#include <thrust/tabulate.h>
+
+#include <unittest/unittest.h>
+
+template <typename ForwardIterator, typename UnaryOperation>
+void tabulate(my_system& system, ForwardIterator, ForwardIterator, UnaryOperation)
+{
+  system.validate_dispatch();
+}
+
+void TestTabulateDispatchExplicit()
+{
+  thrust::device_vector<int> vec(1);
+
+  my_system sys(0);
+  thrust::tabulate(sys, vec.begin(), vec.end(), ::cuda::std::identity{});
+
+  ASSERT_EQUAL(true, sys.is_valid());
+}
+DECLARE_UNITTEST(TestTabulateDispatchExplicit);
+
+template <typename ForwardIterator, typename UnaryOperation>
+void tabulate(my_tag, ForwardIterator first, ForwardIterator, UnaryOperation)
+{
+  *first = 13;
+}
+
+void TestTabulateDispatchImplicit()
+{
+  thrust::device_vector<int> vec(1);
+
+  thrust::tabulate(thrust::retag<my_tag>(vec.begin()), thrust::retag<my_tag>(vec.end()), ::cuda::std::identity{});
+
+  ASSERT_EQUAL(13, vec.front());
+}
+DECLARE_UNITTEST(TestTabulateDispatchImplicit);
+
+template <class Vector>
+void TestTabulateSimple()
+{
+  using namespace thrust::placeholders;
+
+  Vector v(5);
+
+  thrust::tabulate(v.begin(), v.end(), ::cuda::std::identity{});
+
+  Vector ref{0, 1, 2, 3, 4};
+  ASSERT_EQUAL(v, ref);
+
+  thrust::tabulate(v.begin(), v.end(), -_1);
+
+  ref = {0, -1, -2, -3, -4};
+  ASSERT_EQUAL(v, ref);
+
+  thrust::tabulate(v.begin(), v.end(), _1 * _1 * _1);
+
+  ref = {0, 1, 8, 27, 64};
+  ASSERT_EQUAL(v, ref);
+}
+DECLARE_VECTOR_UNITTEST(TestTabulateSimple);
+
+template <typename T>
+void TestTabulate(size_t n)
+{
+  using namespace thrust::placeholders;
+
+  thrust::host_vector<T> h_data(n);
+  thrust::device_vector<T> d_data(n);
+
+  thrust::tabulate(h_data.begin(), h_data.end(), _1 * _1 + 13);
+  thrust::tabulate(d_data.begin(), d_data.end(), _1 * _1 + 13);
+
+  ASSERT_EQUAL(h_data, d_data);
+
+  thrust::tabulate(h_data.begin(), h_data.end(), (_1 - 7) * _1);
+  thrust::tabulate(d_data.begin(), d_data.end(), (_1 - 7) * _1);
+
+  ASSERT_EQUAL(h_data, d_data);
+}
+DECLARE_VARIABLE_UNITTEST(TestTabulate);
+
+template <typename T>
+void TestTabulateToDiscardIterator(size_t n)
+{
+  thrust::tabulate(thrust::discard_iterator<thrust::device_system_tag>(),
+                   thrust::discard_iterator<thrust::device_system_tag>(static_cast<std::ptrdiff_t>(n)),
+                   ::cuda::std::identity{});
+
+  // nothing to check -- just make sure it compiles
+}
+DECLARE_VARIABLE_UNITTEST(TestTabulateToDiscardIterator);