Multiplication speed (of uint8_t) could be improved by directly generating code for the multiplication instead of calling a support function.
When using the fast multiplication code from #1896290, this should even result in code size reduction: The multiplication code itself should be one byte less than the code for calling the support function.
I make this a separate feature request since #1896290 is easier to implement and less likely to break anything (I want to implement #1896290 for 2.8.0. This feature request is for post-2.8.0 sdcc).
Log in to post a comment.