CMake compiles CUDA with a specified computing capability of 8.6, compilation option -- ptxas options=- v, printing memory information issue

When compiling the CUDA program with CMake, specify a computing capability of 8.6 and set CMAKE_ CUDA_ FLAGS is - arch=sm_ 86 – ptxas options=- v, why does the printed memory information include the Compiling entry function_ Z10AddFromGPUPfS_ S_ I ‘for’ sm_ 52 '.
CMake:
cmake_minimum_required(VERSION 3.20)

project(Computing_resource_allocation LANGUAGES CXX CUDA)
set(CMAKE_EXPORT_COMPILE_COMMANDS ON)
set(CMAKE_CUDA_STANDARD 14)
set(CMAKE_CUDA_STANDARD_REQUIRED ON)
set(CMAKE_CXX_STANDARD 14)
set(CMAKE_CXX_STANDARD_REQUIRED ON)

set(CMAKE_CUDA_FLAGS “${CMAKE_CUDA_FLAGS} -O3 -arch=sm_86 --ptxas-options=-v”)

set(CMAKE_CUDA_FLAGS_DEBUG=“-G -g -O0”)

find_package(CUDA 11.7 REQUIRED)
if (CUDA_FOUND)
message(STATUS “Found CUDA ${CUDA_VERSION_STRING} at ${CUDA_TOOLKIT_ROOT_DIR}”)
message(STATUS “CUDA_INCLUDE_DIRS: ${CUDA_INCLUDE_DIRS}”)
message(STATUS “CUDA_LIBRARIES: ${CUDA_LIBRARIES}”)
message(STATUS “CUDA_LIBRARY_DIRS: ${CUDA_LIBRARY_DIRS}”)
else()
message(FATAL_ERROR “Cannot find CUDA”)
endif()

add_executable(arrayAddition arrayAddition.cu common.cuh)
target_include_directories(arrayAddition PRIVATE ${CUDA_INCLUDE_DIRS})
target_link_libraries(arrayAddition PRIVATE ${CUDA_LIBRARIES})

Execute make:
[ 50%] Building CUDA object CMakeFiles/arrayAddition.dir/arrayAddition.cu.o
ptxas info : 0 bytes gmem
ptxas info : Compiling entry function ‘_Z10addFromGPUPfS_S_i’ for ‘sm_86’
ptxas info : Function properties for _Z10addFromGPUPfS_S_i
0 bytes stack frame, 0 bytes spill stores, 0 bytes spill loads
ptxas info : Used 12 registers, 380 bytes cmem[0]
ptxas info : 0 bytes gmem
ptxas info : Compiling entry function ‘_Z10addFromGPUPfS_S_i’ for ‘sm_52’
ptxas info : Function properties for _Z10addFromGPUPfS_S_i
0 bytes stack frame, 0 bytes spill stores, 0 bytes spill loads
ptxas info : Used 8 registers, 348 bytes cmem[0]
[100%] Linking CUDA executable arrayAddition
[100%] Built target arrayAddition

I am using CMAKE_ CUDA_ FLAGS specifies - arch=sm_ 86, why is there still a Compiling entry function_ Z10AddFromGPUPfS_ S_ I ‘for’ sm_ 52 '.
When compiling with When compiling with a, there will be no occurrence of b., there will be no occurrence of ‘_Z10addFromGPUPfS_S_i’ for ‘sm_52’.

Why is this?