[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/mr_new.cu

@@ -0,0 +1,37 @@
+#include <thrust/fill.h>
+#include <thrust/mr/new.h>
+
+#include <unittest/unittest.h>
+
+template <typename MemoryResource>
+void TestAlignment(MemoryResource memres, std::size_t size, std::size_t alignment)
+{
+  void* ptr = memres.do_allocate(size, alignment);
+  ASSERT_EQUAL(reinterpret_cast<std::size_t>(ptr) % alignment, 0u);
+
+  char* char_ptr = reinterpret_cast<char*>(ptr);
+  thrust::fill(char_ptr, char_ptr + size, char{});
+
+  memres.do_deallocate(ptr, size, alignment);
+}
+
+static const std::size_t MinTestedSize  = 32;
+static const std::size_t MaxTestedSize  = 8 * 1024;
+static const std::size_t TestedSizeStep = 1;
+
+static const std::size_t MinTestedAlignment   = 16;
+static const std::size_t MaxTestedAlignment   = 4 * 1024;
+static const std::size_t TestedAlignmentShift = 1;
+
+void TestNewDeleteResourceAlignedAllocation()
+{
+  for (std::size_t size = MinTestedSize; size <= MaxTestedSize; size += TestedSizeStep)
+  {
+    for (std::size_t alignment = MinTestedAlignment; alignment <= MaxTestedAlignment;
+         alignment <<= TestedAlignmentShift)
+    {
+      TestAlignment(thrust::mr::new_delete_resource(), size, alignment);
+    }
+  }
+}
+DECLARE_UNITTEST(TestNewDeleteResourceAlignedAllocation);