[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/examples/stf/void_data_interface.cu

@@ -0,0 +1,57 @@
+//===----------------------------------------------------------------------===//
+//
+// 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 Illustrate how to use the void data interface
+ *
+ */
+
+#include <cuda/experimental/stf.cuh>
+
+using namespace cuda::experimental::stf;
+
+__global__ void dummy_kernel() {}
+
+int main()
+{
+  context ctx;
+
+  auto token = ctx.token();
+  ctx.task(token.write())->*[](cudaStream_t) {
+
+  };
+
+  void_interface sync;
+  auto token2 = ctx.logical_data(sync);
+
+  auto token3 = ctx.token();
+  ctx.task(token2.write(), token.read())->*[](cudaStream_t) {
+
+  };
+
+  // Do not pass useless arguments by removing void_interface arguments
+  // Note that the rw() access is possible even if there was no prior write()
+  // or actual underlying data.
+  ctx.task(token3.rw(), token.read())->*[](cudaStream_t) {
+
+  };
+
+  ctx.cuda_kernel(token3.rw())->*[]() {
+    return cuda_kernel_desc{dummy_kernel, 16, 128, 0};
+  };
+
+  EXPECT(token.is_void_interface());
+  EXPECT(token2.is_void_interface());
+  EXPECT(token3.is_void_interface());
+
+  ctx.finalize();
+}