HT helps a P4, because it has such a long pipeline, some data doesnt get to the end in time, and must start over, leaving extra room in the pipe. HT prepares data to take that place if the first set of data goes back, so that no matter what, the entire pipeline of the cpu is in use, therefore being much more efficient.
HT would not benefit that athlon 64 at all, as it has a very short, and super efficient pipeline system, and data that gets sent back(if any) will be back where it was before any SMT technology could replace if with something else. That is why amd doesnt use it, nor intend to any time soon.