Re-check Poseidon Parameters

[DanieleDiBenedetto]

Generation scripts of Poseidon parameters at https://extgit.iaik.tugraz.at/krypto/hadeshash/-/tree/master have been recently updated.
We should re-execute the scripts for our curves and check if any of the parameters we currently have in the code has been affected by the changes.

[UlrichHaboeck75]

The new script is already applied for the new parameters in the branch [UpdatedPoseidonParameters] (https://github.com/HorizenOfficial/ginger-lib/tree/UpdatedPoseidonParameters).
However, when applying it to the 255 bit field sized Tweedles, the outcome of the calc_round_numbers.py script slightly differs from the recommendations in the Poseidon paper: for the x^5 poseidon with t=3 field elements and a security level of M=128, the script outputs R_F= 8 and R_p=56 (including security margin), in contrast to R_F= 8 and R_p =57 as recommended in the paper.

[UlrichHaboeck75]

According to the Poseidon authors, the round numbers from the paper increase the number of partial rounds to be a multiple of t=3 (for a compact circuit, see Appendix E ). As long as we do not use this strategy for the circuit, we can keep with R_p=56 rounds.

[UlrichHaboeck75]

We need to evaluate how the circuit strategy from Appendix E relates to the issue of propagation of linear combination.
If it does better than separate propagation stopping contstraints, we should again check whether it is even better for us to take R_p=57.