Add all rounding modes

### Motivation

We only support truncation and rounding-nearest-even in our logic and value side of floating-point.
We have enumerated the different rounding modes but not implemented them. If we have more components like RoundRNE and we factor out the value-side equivalents, we will have a full bevy of rounders available for our FP components.

We need to refactor our components to also use these rounders, which requires setting up the mantissas for the rounders to perform.

### Desired solution

Create value-side rounders and refactor the ofDouble routine to use them.
Create logic-side rounders and refactor the adders and multipliers and converters to use them.
Write tests that, at the very least, compare the value side and logic side implementations using exhaustive tests.

### Alternatives considered

_No response_

### Additional details

_No response_