[opus] Speed performance of 1.0.3

[Sergey G]

Hi Opus team,

For some reasons we have to use opus 1.0.3 fixed point implementation. There is an information about significant improvement of performance in opus 1.1-beta. Could you please share information about which of the changes in 1.1-beta could improve the performance without loosing bit -to-bit compatibility to 1.0.3 ?

For instance, cwrsi() from 1.1-beta can improve  the speed performance and generates the same bit exact result as cwrsi() from 1.0.3.

Do you have a list of such functions ?

Thanks,
SG