[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/execution/test_visit.cu

@@ -0,0 +1,81 @@
+//===----------------------------------------------------------------------===//
+//
+// 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.
+//
+//===----------------------------------------------------------------------===//
+
+#include <cuda/std/__algorithm/max.h>
+
+#include <cuda/experimental/execution.cuh>
+
+#include "testing.cuh" // IWYU pragma: keep
+
+namespace
+{
+struct S0
+{};
+struct S1
+{
+  int a;
+};
+struct S2
+{
+  int a, b;
+};
+static_assert(cudax_async::structured_binding_size<S0> == 0);
+static_assert(cudax_async::structured_binding_size<S1> == 1);
+static_assert(cudax_async::structured_binding_size<S2> == 2);
+
+template <class Fn>
+struct recursive_lambda
+{
+  template <class... Args>
+  auto operator()(Args&&... args)
+  {
+    return fn(*this, cuda::std::forward<Args>(args)...);
+  }
+
+  Fn fn;
+};
+
+template <class Fn>
+recursive_lambda(Fn) -> recursive_lambda<Fn>;
+
+C2H_TEST("sender visitation API works", "[visit]")
+{
+  int leaves = 0;
+  int depth  = 0;
+
+  auto snd = cudax_async::when_all(
+    cudax_async::just(3), //
+    cudax_async::just(0.1415),
+    cudax_async::then(cudax_async::just(0.1415), [](double f) {
+      return f;
+    }));
+  auto snd1 = std::move(snd) | cudax_async::then([](int x, double y, double z) {
+                return x + y + z;
+              });
+
+  auto count_leaves = recursive_lambda{[](auto& self, int& leaves, auto, auto&, auto&... child) {
+    leaves += (sizeof...(child) == 0);
+    ((cudax_async::visit(self, child, leaves)), ...);
+  }};
+
+  cudax_async::visit(count_leaves, snd1, leaves);
+  CHECK(leaves == 3);
+
+  auto max_depth = recursive_lambda{[i = 0](auto& self, int& depth, auto, auto&, auto&... child) mutable {
+    ++i;
+    depth = cuda::std::max(depth, i);
+    ((cudax_async::visit(self, child, depth)), ...);
+    --i;
+  }};
+
+  cudax_async::visit(max_depth, snd1, depth);
+  CHECK(depth == 4);
+}
+} // namespace