Use "poseidon_hash_2" for merkle trees instead of poseidon_hash_many

[feltroidprime]

The hash trait is using "poseidon_hash_many" or "array hashing" which is more costly and useless if we know in advance how many elements we hash.

A standard way of hashing two elements is calling the permutation on (x, y, 2) and returning the first element of the state.

see

• https://github.com/starkware-libs/cairo-lang/blob/8e11b8cc65ae1d0959328b1b4a40b92df8b58595/src/starkware/cairo/common/poseidon_hash.py#L22-L31

• https://docs.starknet.io/architecture-and-concepts/cryptography/hash-functions/#poseidon_hash

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
raito	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jan 5, 2025 9:37pm

[feltroidprime]

Where is the script generating the jsons tests for utreexo roots ?

[maciejka]

Where is the script generating the jsons tests for utreexo roots ?

It is generated by Floresta node: https://github.com/keep-starknet-strange/raito/blob/main/docs/data.md#utreexo-data.

@m-kus will you have a look?

[m-kus]

Note to future us: we'd need to change the implementation in the accumulator here https://github.com/Davidson-Souza/rpc-utreexo-bridge/blob/115fb5f21a19fc8ed30dab14d2b6558f782fe9b2/src/udata.rs#L514 and reindex the bridge node to recreate test data

[feltroidprime]

Better now than after proving 1TB of data