Small performance optimizations:
- Core:
- 32-bit Core now has properly implemented support for generating Armv8-a 32-bit code instead of being forced to armv6k, thanks to that we can use less instructions to achieve the same thing
- 32-bit Core now properly recognizes alignment of SharedMemory so compiler properly optimizes instructions accessing it instead of taking worst case approach
- NX-FPS:
- Remove most of repetetive code
- Cache more frequently values returned by weakly linked function calls
- Avoid using float divisions in functions that are called multiple times in one frame
- Sysmodule:
- Apply proper fix instead of workaround for SM
- Fix an overflow in internal display pixel clock calculation which resulted in wrong reported refresh rate for overclocked displays (was not an issue in stock SaltyNX because overclocking internal displays is unsupported)