A few benchmarks of Pi:
http://www.roylongbottom.org.uk/Raspber ... hmarks.htm
I don't think there is a real need (and sense) in porting ST to another architecture. From my point of view it's a huge effort even to get something working optimally across various operating systems, using the same hardware platform (and Windows, Linux, IOS - similarly to ST - all run exceptionally well on *typical* Intel x86 CPUs).
Look at these two examples: Omnia 9 is a typical x86 machine (Intel x86/x64 SISC, One of the Core generation CPU, using 4 processing cores), while Optimod FM 5500 is a RISC device (like ARM), with dual core CPU from Freescale Semiconductor. Both do what they are designed to do. And they do it well. But you will never be able to port software from one to another.
I think there would have to be another version of some audio enhancer (much simplier) designed and coded specifically for ARMs (RISC). It would be much easier to code it from the ground up, and it would have a chance to run efficiently enough, so that everyone is satisfied. But this means it would also cost some time to learn the new device at first, then how to write the code for it and finally optimize. Is it worth it?