this commit adds additional binop optimizations to venom. these are
the analog of `optimize_binop` in the legacy optimizer and take us one
step closer to being able to bypass the legacy optimizer in the venom
pipeline.

the rules have been added to the algebraic optimizations pass. adding
them to the SCCP pass was considered, but to keep the rules all in one
place and to keep each pass more self-contained, they were added to the
algebraic optimizations pass, with additional SCCP passes performed for
branch/cfg reduction.

some additional machinery has been added, including `lit_eq` which
compares two IRLiterals in modular arithmetic, and InstructionUpdator,
which updates instructions while also updating the DFG in-place. this
should be useful to use for other passes in the future (although out of
scope for this PR).

additional passes have been added since they were found to improve
codesize. in particular, running AlgebraicOptimization multiple times,
and running SCCP after LoadElimination, seem to have some additional
benefits.