Hmmmmm that's not really a question that can be answered without restating it.

If AMD could make a 3.2GHz Athlon with HyperThreading it would outperform anything Intel has on the desktop right now. AMD's IPC is enough higher that clock speed parity would not be a good thing for Intel.

However, the current Athlons wouldn't do 3.2GHz because that's faster than that architecture can be pushed. Most designs end up having hard ceilings (much like the Coppermine PIII which could never really be made reliable at 1.13GHz) and 3.2GHz is above the Athlons. There may be ways to get one up to 3GHz with high voltages and exotic cooling but it's unlikely to be practical. As to Hyperthreading, AMD could implement it, they not only have extensive cross-licensing deals with Intel but also have there own patents in the area. However, there would be much less point doing it on an Athlon than a P4.

The P4 has a very long pipeline which is subject to stalls, which is what happens when you get a bubble with no data in it passing down the pipeline. Remember if the CPU isn't fed with data properly that empty clock cycle will propagate all the way down the pipeline, and it takes a P4 20 cycles to get data from one end of the pipe to the other. Hyperthreading is a way of having two threads going so if you get a bubble it just feeds data from the other thread to keep the pipelines full. Now the Athlon has shorter pipes and less trouble keeping them full, so it woudn't benefit as much from Hyperthreading as a P4.

Also the Athlon has more execution units than P4 so it can do more in parallel and catch up from a stall more quickly. Anyway, the end result is the way to get good performance from a P4 is to keep the pipes full all the time, and hyperthreading is one way to help do that.

This is partially why AMD did implement SSE2 on the Athlon64 and Opteron but not hyperthreading, they could take more advantage from one than the other.