Has anyone tested changing the memory order on the load()/store() of their parameter values? As I understand it, you might see some difference in performance (/horrible bugs?) on ARM.
It influences the re-ordering the compiler is allowed to do when optimizing (moving inside or outside of a locked scope).
Was a good talk on ADC addressing this.
But I don’t think you are going to get much of a performance boost on parameter loading. Horrible bugs might be more likely…