sharing pinned memory between GPUs/contexts

With the new release, is there yet a way to use pinned memory between two contexts in two separate threads?

Thanks,
Brian

The early access CUDA 2.0 is CUDA 1.1 plus double precision.
There are no new features.