Optimize Cardano transaction prover performances with parallelization

[jpraynaud]

Why

We would like to parallelize some parts of the prover in order to leverage parallelization on some tasks.

What

Add parallelization for retrieval/computation of the Merkle trees of the block ranges and for the computation of their poofs with rayon.

How

• Implement par_iter in Merkle tree computation in the prover
• Implement par_iter in Merkle map proof computation
• Benchmark the performances improvements

[jpraynaud]

The benchmarks show slight performance degradation. The PR will not be merged and the issue is closed.