[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/cub/cmake/CubBuildCompilerTargets.cmake

@@ -0,0 +1,56 @@
+# This file provides the following function which defines the following targets:
+#
+# cub.compiler_interface
+# - Interface target that includes all compiler settings for cub tests, etc.
+
+function(cub_build_compiler_targets)
+  cccl_get_cub()
+  cccl_get_libcudacxx()
+  cccl_get_thrust()
+
+  thrust_create_target(cub.thrust HOST CPP DEVICE CUDA)
+
+  set(ptxas_compile_options)
+  if (CCCL_ENABLE_PTXAS_WARNINGS)
+    list(
+      APPEND ptxas_compile_options
+      "--warn-on-spills"
+      "--warn-on-local-memory-usage"
+    )
+  endif()
+
+  set(cuda_compile_options)
+  set(cxx_compile_options)
+  set(cxx_compile_definitions)
+
+  # append ptxas compile options to cuda_compile_options with compiler specific prefix
+  foreach (ptxas_compile_option ${ptxas_compile_options})
+    if (
+      "${CMAKE_CUDA_COMPILER_ID}" STREQUAL "NVIDIA"
+      OR "${CMAKE_CUDA_COMPILER_ID}" STREQUAL "NVHPC"
+    )
+      list(APPEND cuda_compile_options "-Xptxas=${ptxas_compile_option}")
+    elseif ("${CMAKE_CUDA_COMPILER_ID}" STREQUAL "CLANG")
+      list(APPEND cuda_compile_options "-Xcuda-ptxas ${ptxas_compile_option}")
+    endif()
+  endforeach()
+
+  cccl_build_compiler_interface(
+    cub.compiler_flags
+    "${cuda_compile_options}"
+    "${cxx_compile_options}"
+    "${cxx_compile_definitions}"
+  )
+
+  add_library(cub.compiler_interface INTERFACE)
+  target_link_libraries(
+    cub.compiler_interface
+    INTERFACE
+      # order matters here, we need the project options to override the cccl options.
+      cccl.compiler_interface
+      cub.compiler_flags
+      libcudacxx::libcudacxx
+      CUB::CUB
+      cub.thrust
+  )
+endfunction()