Not sure. According to nVidia's OpenCL Programming Guide for the CUDA Architecture, Appendix A.3:
All compute devices supports the cl_khr_byte_addressable_store extension.
Devices of compute capability 1.1 and higher support the cl_khr_global_int32_base_atomics, cl_khr_global_int32_extended_atomics, cl_khr_local_int32_base_atomics, and cl_khr_local_int32_extended_atomics extensions.
Does this answer your question, I hope?