[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/libcudacxx/include/cuda/__utility/immovable.h

@@ -0,0 +1,50 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of libcu++, the C++ Standard Library for your entire system,
+// under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+// SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef _CUDA___UTILITY_IMMOVABLE_H
+#define _CUDA___UTILITY_IMMOVABLE_H
+
+#include <cuda/std/detail/__config>
+
+#if defined(_CCCL_IMPLICIT_SYSTEM_HEADER_GCC)
+#  pragma GCC system_header
+#elif defined(_CCCL_IMPLICIT_SYSTEM_HEADER_CLANG)
+#  pragma clang system_header
+#elif defined(_CCCL_IMPLICIT_SYSTEM_HEADER_MSVC)
+#  pragma system_header
+#endif // no system header
+
+#include <cuda/std/__cccl/prologue.h>
+
+_CCCL_BEGIN_NAMESPACE_CUDA
+
+// BUG (gcc#98995): copy elision fails when initializing a [[no_unique_address]] field
+// from a function returning an object of class type by value.
+// See: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98995
+#if _CCCL_COMPILER(GCC)
+// By declaring the move constructor but not defining it, any TU that ODR-uses the move
+// constructor will cause a linker error.
+#  define _CCCL_IMMOVABLE(_XP) _CCCL_API _XP(_XP&&) noexcept
+#else // ^^^ _CCCL_COMPILER(GCC) ^^^ / vvv !_CCCL_COMPILER(GCC) vvv
+#  define _CCCL_IMMOVABLE(_XP) _XP(_XP&&) = delete
+#endif // !_CCCL_COMPILER(GCC)
+
+// Classes can inherit from this type to become immovable.
+struct _CCCL_TYPE_VISIBILITY_DEFAULT __immovable
+{
+  _CCCL_HIDE_FROM_ABI __immovable() = default;
+  _CCCL_IMMOVABLE(__immovable);
+};
+
+_CCCL_END_NAMESPACE_CUDA
+
+#include <cuda/std/__cccl/epilogue.h>
+
+#endif // _CUDA___UTILITY_IMMOVABLE_H