[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/examples/stf/mandelbrot.cu

@@ -1,129 +0,0 @@
-//===----------------------------------------------------------------------===//
-//
-// Part of CUDASTF 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) 2022-2024 NVIDIA CORPORATION & AFFILIATES.
-//
-//===----------------------------------------------------------------------===//
-
-/**
- * @file
- * @brief A transparent multi-GPU implementation of Mandelbrot fractal using parallel_for
- */
-
-#include <cuda/experimental/stf.cuh>
-
-#include <fstream>
-#include <iostream>
-
-using namespace cuda::experimental::stf;
-
-int main(int argc, char** argv)
-{
-  context ctx;
-
-  // Image dimensions
-  size_t width  = 2000;
-  size_t height = 1000;
-
-  // Complex plane boundaries
-  double xMin = -2.0;
-  double xMax = 1.0;
-  double yMin = -1.5;
-  double yMax = 1.5;
-
-  // Maximum number of iterations
-  int maxIterations = 256;
-
-  // Describe a 2D array of integers of size (width x height)
-  auto lbuffer = ctx.logical_data(shape_of<slice<int, 2>>(width, height));
-
-  cudaEvent_t start, stop;
-  cuda_safe_call(cudaEventCreate(&start));
-  cuda_safe_call(cudaEventCreate(&stop));
-
-  cuda_safe_call(cudaEventRecord(start, ctx.fence()));
-
-  // Compute each pixel
-  ctx.parallel_for(blocked_partition(), exec_place::all_devices(), lbuffer.shape(), lbuffer.write())
-      ->*[=] _CCCL_DEVICE(size_t x, size_t y, auto buffer) {
-            // Map pixel coordinates to complex plane
-            // c = cr + i ci
-            double cr = x * (xMax - xMin) / width + xMin;
-            double ci = y * (yMax - yMin) / height + yMin;
-
-            // z = zr + i zi
-            double zr = 0.0;
-            double zi = 0.0;
-
-            int iterations = 0;
-
-            // Evaluate depth
-            while (zr * zr + zi * zi < 4 && iterations < maxIterations)
-            {
-              // compute : z = z * z + c;
-              //
-              // z = (zr + i zi) (zr + i zi) + cr + i ci
-              // z = zr zr - zi zi + 2 i zi zr + cr + i ci
-              // zr = (zr zr - zi zi + cr)
-              // zi = (2 zi zr + ci)
-              double zr_prev = zr;
-              double zi_prev = zi;
-              zr             = zr_prev * zr_prev - zi_prev * zi_prev + cr;
-              zi             = 2.0 * zr_prev * zi_prev + ci;
-
-              iterations++;
-            }
-
-            buffer(x, y) = iterations;
-          };
-
-  cuda_safe_call(cudaEventRecord(stop, ctx.fence()));
-
-  if (argc > 1)
-  {
-    auto fileName = std::string(argv[1]);
-
-    // Generate a PPM file from the buffer
-    ctx.host_launch(lbuffer.read())->*[&](auto buffer) {
-      std::ofstream imageFile(fileName, std::ios::binary);
-      if (!imageFile)
-      {
-        std::cerr << "Failed to create image file: " << fileName << '\n';
-        return;
-      }
-
-      imageFile << "P6\n";
-      imageFile << width << " " << height << "\n";
-      imageFile << "255\n";
-
-      for (size_t y = 0; y < height; y++)
-      {
-        for (size_t x = 0; x < width; x++)
-        {
-          int iterations = buffer(x, y);
-          // Convert iterations to RGB values
-          unsigned char r = (iterations % 8) * 32;
-          unsigned char g = (iterations % 16) * 16;
-          unsigned char b = (iterations % 32) * 8;
-
-          // Write pixel data to file
-          imageFile << r << g << b;
-        }
-      }
-
-      imageFile.close();
-      std::cout << "Mandelbrot image generated and saved as " << fileName << '\n';
-    };
-  }
-
-  ctx.finalize();
-
-  // Must call this first, see e.g.
-  // https://stackoverflow.com/questions/6551121/cuda-cudaeventelapsedtime-returns-device-not-ready-error
-  cuda_safe_call(cudaEventSynchronize(stop));
-
-  fprintf(stderr, "Mandelbrot took %.2f ms\n", cuda_try<cudaEventElapsedTime>(start, stop));
-}