CUDA 2.0 and 64bit data types

From reading the FAQ and other postings in this forum, I understand that 64bit double precision floating point is not supported due to current generation GPU limitation (Geforce 8x series).

But what about long long and unsigned long long data type? Are these two data types supported?