feat: CCCL CachingDeviceAllocator preload — 完整依赖链 288 files

从 cccl_upstream 递归追踪 cub/util_allocator.cuh 的全部 include 依赖:
  cub/         9 files (config, util_*, version, detect_cuda_runtime)
  cuda/        libcudacxx type_traits, concepts, algorithm, iterator...
  nv/          target macros, preprocessor

总计 288 个头文件 (1.4MB)，打包到 include/ 目录，编译时 -I include
即可完全脱离 CCCL 原始目录结构。

.cu 文件直接 #include <cub/util_allocator.cuh>，
走原版 CUB CachingDeviceAllocator，零 mock。

BI-V100 参数: growth=2 bins=[8..32] max_cached=8GB/device

qwen3_6_scripts/cccl_preload/include/cub/config.cuh

@@ -0,0 +1,29 @@
+// SPDX-FileCopyrightText: Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.
+// SPDX-License-Identifier: BSD-3
+
+/**
+ * \file
+ * Static configuration header for the CUB project.
+ */
+
+#pragma once
+
+// For _CCCL_IMPLICIT_SYSTEM_HEADER
+#include <cuda/__cccl_config> // IWYU pragma: export
+
+#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 <cub/util_arch.cuh> // IWYU pragma: export
+#include <cub/util_cpp_dialect.cuh> // IWYU pragma: export
+#include <cub/util_macro.cuh> // IWYU pragma: export
+#include <cub/util_namespace.cuh> // IWYU pragma: export
+
+#if !_CCCL_COMPILER(NVRTC)
+#  include <cuda/__nvtx/nvtx.h>
+#endif // !_CCCL_COMPILER(NVRTC)