[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/thrust/testing/decompose.cu

@@ -0,0 +1,84 @@
+#include <thrust/system/detail/internal/decompose.h>
+
+#include <unittest/unittest.h>
+
+using thrust::system::detail::internal::uniform_decomposition;
+
+void TestUniformDecomposition()
+{
+  {
+    uniform_decomposition<int> ud(10, 10, 1);
+
+    // [0,10)
+    ASSERT_EQUAL(ud.size(), 1);
+    ASSERT_EQUAL(ud[0].begin(), 0);
+    ASSERT_EQUAL(ud[0].end(), 10);
+    ASSERT_EQUAL(ud[0].size(), 10);
+  }
+
+  {
+    uniform_decomposition<int> ud(10, 20, 1);
+
+    // [0,10)
+    ASSERT_EQUAL(ud.size(), 1);
+    ASSERT_EQUAL(ud[0].begin(), 0);
+    ASSERT_EQUAL(ud[0].end(), 10);
+    ASSERT_EQUAL(ud[0].size(), 10);
+  }
+
+  {
+    uniform_decomposition<int> ud(8, 5, 2);
+
+    // [0,5)[5,8)
+    ASSERT_EQUAL(ud.size(), 2);
+    ASSERT_EQUAL(ud[0].begin(), 0);
+    ASSERT_EQUAL(ud[0].end(), 5);
+    ASSERT_EQUAL(ud[0].size(), 5);
+    ASSERT_EQUAL(ud[1].begin(), 5);
+    ASSERT_EQUAL(ud[1].end(), 8);
+    ASSERT_EQUAL(ud[1].size(), 3);
+  }
+
+  {
+    uniform_decomposition<int> ud(8, 5, 3);
+
+    // [0,5)[5,8)
+    ASSERT_EQUAL(ud.size(), 2);
+    ASSERT_EQUAL(ud[0].begin(), 0);
+    ASSERT_EQUAL(ud[0].end(), 5);
+    ASSERT_EQUAL(ud[0].size(), 5);
+    ASSERT_EQUAL(ud[1].begin(), 5);
+    ASSERT_EQUAL(ud[1].end(), 8);
+    ASSERT_EQUAL(ud[1].size(), 3);
+  }
+
+  {
+    uniform_decomposition<int> ud(10, 1, 2);
+
+    // [0,5)[5,10)
+    ASSERT_EQUAL(ud.size(), 2);
+    ASSERT_EQUAL(ud[0].begin(), 0);
+    ASSERT_EQUAL(ud[0].end(), 5);
+    ASSERT_EQUAL(ud[0].size(), 5);
+    ASSERT_EQUAL(ud[1].begin(), 5);
+    ASSERT_EQUAL(ud[1].end(), 10);
+    ASSERT_EQUAL(ud[1].size(), 5);
+  }
+
+  {
+    // [0,4)[4,8)[8,10)
+    uniform_decomposition<int> ud(10, 2, 3);
+
+    ASSERT_EQUAL(ud.size(), 3);
+    ASSERT_EQUAL(ud[0].begin(), 0);
+    ASSERT_EQUAL(ud[0].end(), 4);
+    ASSERT_EQUAL(ud[0].size(), 4);
+    ASSERT_EQUAL(ud[1].begin(), 4);
+    ASSERT_EQUAL(ud[1].end(), 8);
+    ASSERT_EQUAL(ud[1].size(), 4);
+    ASSERT_EQUAL(ud[2].begin(), 8);
+    ASSERT_EQUAL(ud[2].end(), 10);
+    ASSERT_EQUAL(ud[2].size(), 2);
+  }
+}
+DECLARE_UNITTEST(TestUniformDecomposition);