pub fn operand_bits(p: Precision) -> Option<u32>
Operand bit width for the integer reference (None for float operands).
None