[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/cudax/test/stf/places/place_partition.cu

@@ -0,0 +1,53 @@
+//===----------------------------------------------------------------------===//
+//
+// 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.
+//
+//===----------------------------------------------------------------------===//
+
+#include <cuda/experimental/__places/place_partition.cuh>
+#include <cuda/experimental/__stf/internal/stf_places_partition_into_stf.cuh>
+#include <cuda/experimental/stf.cuh>
+
+using namespace cuda::experimental::stf;
+
+void print_partition(async_resources_handle& handle, exec_place place, place_partition_scope scope)
+{
+  fprintf(stderr, "-----------\n");
+  fprintf(
+    stderr, "PARTITION %s (scope: %s):\n", place.to_string().c_str(), place_partition_scope_to_string(scope).c_str());
+  for (auto sub_place : place_partition(place, handle, scope))
+  {
+    fprintf(stderr, "[%s] subplace: %s\n", place.to_string().c_str(), sub_place.to_string().c_str());
+  }
+
+  fprintf(stderr, "-----------\n");
+}
+
+int main()
+{
+#if _CCCL_CTK_BELOW(12, 4)
+  fprintf(stderr, "Green contexts are not supported by this version of CUDA: skipping test.\n");
+  return 0;
+#else // ^^^ _CCCL_CTK_BELOW(12, 4) ^^^ / vvv _CCCL_CTK_AT_LEAST(12, 4) vvv
+  async_resources_handle handle;
+
+  print_partition(handle, exec_place::all_devices(), place_partition_scope::cuda_device);
+
+  print_partition(handle, exec_place::all_devices(), place_partition_scope::cuda_stream);
+
+  print_partition(handle, exec_place::current_device(), place_partition_scope::cuda_stream);
+
+  print_partition(handle, exec_place::current_device(), place_partition_scope::green_context);
+  print_partition(handle, exec_place::current_device(), place_partition_scope::green_context);
+
+  print_partition(handle, exec_place::repeat(exec_place::current_device(), 4), place_partition_scope::green_context);
+
+  print_partition(handle, exec_place::current_device(), place_partition_scope::cuda_device);
+
+  print_partition(handle, exec_place::repeat(exec_place::current_device(), 4), place_partition_scope::cuda_stream);
+#endif // ^^^ _CCCL_CTK_AT_LEAST(12, 4) ^^^
+}