* Make frame processing hashing parallel
* Do not build address for new style proofs
* Validate mint requests in parallel
* Do not recompute byte form of address in mint
* Remove unused output
* Use single append
* Cache verify challenge proof results
