3 ms·
> you can only multiply two 16 bit numbers without fear of loss of precision (without doing a fair amount of gymnastics) You can do signed and unsigned 32-bit
by TazeTSchnitzel 11y ago
> you can only multiply two 16 bit numbers without fear of loss of precision (without doing a fair amount of gymnastics)
You can do signed and unsigned 32-bit integer multiplication, but there's modulo/overflow there.
- Franciscouzo 11y agoYou can use 26 bits integers for multiplication without fear of overflowing with IEEE 754 doubles.