You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Filecoin has an optimized implementation of Poseidon that has gone through an audit by the ADBK Consulting (by Mikhail Vladimirov and Dmitry Khovratovich)
It seems to include two changes: (1) treatment of the round constants and (2) the use of sparse MDS matrices.
Neptune is fortunately MIT/Apache2.
Problem Definition
It may be interesting to look at their implementations and see if it is compatible with the current implementation, and whether or not it can bring performance improvement.
Proposal
Analysis the compatibility and decide whether or not to implement it.
For Admin Use
Not duplicate issue
Appropriate labels applied
Appropriate contributors tagged
Contributor assigned/self-assigned
The text was updated successfully, but these errors were encountered:
We're okay with "sponges" that do not support converting from sponge back into absorb to give runtime errors? I'd think runtime errors would be the only option for something like say a sponge that permits converting back to absorb only if two field elements remain unsqueezed.
Summary
Filecoin has an optimized implementation of Poseidon that has gone through an audit by the ADBK Consulting (by Mikhail Vladimirov and Dmitry Khovratovich)
https://github.com/filecoin-project/neptune/blob/master/spec/poseidon_spec.pdf
It seems to include two changes: (1) treatment of the round constants and (2) the use of sparse MDS matrices.
Neptune is fortunately MIT/Apache2.
Problem Definition
It may be interesting to look at their implementations and see if it is compatible with the current implementation, and whether or not it can bring performance improvement.
Proposal
Analysis the compatibility and decide whether or not to implement it.
For Admin Use
The text was updated successfully, but these errors were encountered: