Verkle ipa multiproof is now internally consistent

[mratsim]

This is a followup to the Verkle refactoring from

• Ethereum Verkle IPA refactoring part 2 #397

While simple proof generation / verification had an off-by-one found right away #403,
multiproofs had:

• a multiplication missing in polynomial division that was not tested by the multiproof consistency check add: multiproof consistency test #424
• a zero init missing

The current code is self-consistent, but we need more assurance that the outputs are the same as the other libraries.
Hence we keep #396 open to make sure test coverage is improved and corner cases are hardened.