[CCCL] 瘦身 + 补全: 移除 cudax/python/libcudacxx-tests 冗余文件, 新增 c2h 测试助手 + cmake 构建系统 + 8 个 CUDA thrust examples

变更摘要:
- 删除: cudax/ (783 files, 7.2M) — 实验性组件，竞赛不需要
- 删除: python/ (226 files, 2.0M) — Python 绑定，竞赛不需要
- 删除: libcudacxx/{test,benchmarks,codegen,cmake,share} (4432 files, 31M)
  保留: libcudacxx/include/ (1463 headers, cuda::std 编译依赖)
- 新增: c2h/ (27 files) — CUB Catch2 测试辅助头文件，编译 243 个测试必需
- 新增: cmake/ (29 files) — CCCL 原生 CMake 构建系统
- 新增: thrust/examples/cuda/ (7 files) + cpp_integration/ (1 file)
  async_reduce, custom_temporary_allocation, explicit_cuda_stream,
  global_device_vector, range_view, unwrap_pointer, wrap_pointer, device

结果: cccl_upstream 从 74M→35M (瘦身 53%), 核心内容 100% 保留:
  27/27 tuning headers, 78 benchmarks, 243 tests,
  60 thrust examples, 18 CUB examples, 全部编译头文件

cccl_upstream/cudax/test/algorithm/common.cuh

@@ -1,128 +0,0 @@
-//===----------------------------------------------------------------------===//
-//
-// Part of CUDA Experimental in CUDA C++ Core Libraries,
-// 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) 2024 NVIDIA CORPORATION & AFFILIATES.
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef __ALGORITHM_COMMON__
-#define __ALGORITHM_COMMON__
-
-#include <cuda/algorithm>
-#include <cuda/memory_pool>
-#include <cuda/memory_resource>
-#include <cuda/std/mdspan>
-
-#include <cuda/experimental/container.cuh>
-#include <cuda/experimental/memory_resource.cuh>
-
-#include <testing.cuh>
-#include <utility.cuh>
-
-inline constexpr uint8_t fill_byte    = 1;
-inline constexpr uint32_t buffer_size = 42;
-
-inline int get_expected_value(uint8_t pattern_byte)
-{
-  int result;
-  memset(&result, pattern_byte, sizeof(int));
-  return result;
-}
-
-template <typename Result>
-void check_result_and_erase(cudax::stream_ref stream, Result&& result, uint8_t pattern_byte = fill_byte)
-{
-  int expected = get_expected_value(pattern_byte);
-
-  stream.sync();
-  for (int& i : result)
-  {
-    REQUIRE(i == expected);
-    i = 0;
-  }
-}
-
-template <typename Layout = cuda::std::layout_right, typename Extents>
-auto make_buffer_for_mdspan(Extents extents, char value = 0)
-{
-  cuda::mr::legacy_pinned_memory_resource host_resource;
-  auto mapping = typename Layout::template mapping<decltype(extents)>{extents};
-
-  cudax::uninitialized_buffer<int, cuda::mr::host_accessible> buffer(host_resource, mapping.required_span_size());
-
-  memset(buffer.data(), value, buffer.size_bytes());
-
-  return buffer;
-}
-
-inline auto create_fake_strided_mdspan()
-{
-  cuda::std::dextents<size_t, 3> dynamic_extents{1, 2, 3};
-  cuda::std::array<size_t, 3> strides{12, 4, 1};
-#if _CCCL_CUDA_COMPILER(NVCC, <, 12, 6)
-  auto map = cuda::std::layout_stride::mapping{dynamic_extents, strides};
-#else // ^^^ _CCCL_CUDA_COMPILER(NVCC, <, 12, 6) ^^^ / vvv _CCCL_CUDA_COMPILER(NVCC, >=, 12, 6) vvv
-  cuda::std::layout_stride::mapping map{dynamic_extents, strides};
-#endif // ^^^ _CCCL_CUDA_COMPILER(NVCC, >=, 12, 6) ^^^
-  return cuda::std::mdspan<int, decltype(dynamic_extents), cuda::std::layout_stride>(nullptr, map);
-};
-
-namespace cuda::experimental
-{
-// Need a type that goes through all launch_transform steps, but is not a contiguous_range
-template <typename RelocatableValue = cuda::std::span<int>>
-struct weird_buffer
-{
-  cuda::mr::legacy_pinned_memory_resource& resource;
-  int* data;
-  std::size_t size;
-
-  weird_buffer(::cuda::mr::legacy_pinned_memory_resource& res, std::size_t s)
-      : resource(res)
-      , data((int*) res.allocate_sync(s * sizeof(int)))
-      , size(s)
-  {
-    memset(data, 0, size);
-  }
-
-  ~weird_buffer()
-  {
-    resource.deallocate_sync(data, size);
-  }
-
-  weird_buffer(const weird_buffer&) = delete;
-  weird_buffer(weird_buffer&&)      = delete;
-
-  struct transform_result
-  {
-    int* data;
-    std::size_t size;
-
-    RelocatableValue transformed_argument()
-    {
-      return *this;
-    };
-
-    operator cuda::std::span<int>()
-    {
-      return {data, size};
-    }
-
-    template <typename Extents>
-    operator cuda::std::mdspan<int, Extents>()
-    {
-      return cuda::std::mdspan<int, Extents>{data};
-    }
-  };
-
-  [[nodiscard]] friend transform_result transform_launch_argument(cuda::stream_ref, const weird_buffer& self) noexcept
-  {
-    return {self.data, self.size};
-  }
-};
-} // namespace cuda::experimental
-
-#endif // __ALGORITHM_COMMON__