Consider using a more optimal way to combine the limbs in the btiwise chiplet constraints

Again, an addition chain here might be more efficient. While this description is much cleaner, it might be possible to optimize this in the IR?

_Originally posted by @adr1anh in https://github.com/0xMiden/air-script/pull/466#discussion_r2382324931_

The issue references a discussion in pull request #466 about optimizing limb combination in the bitwise chiplet constraints. Start by reviewing that PR discussion to understand the context. Then examine the relevant IR code for the bitwise chiplet to see how limbs are currently combined. Look for potential addition chain optimizations in the constraint generation.

Scritto dal modello di indicizzazione a partire dal testo della issue.