pub fn precision_int_bits(p: Precision) -> Option<u32>
Bit width of an integer precision (the operand truncation the systolic reference applies).