Innokentii Sennovskii
Innokentii Sennovskii
Stdlib transcript in ultra hashes values under native poseidon and later simply creates a new in-circuit field element from that value. There are no poseidon gates.
Currently we just add constraints and return new elements when we normalize field elements and reduce bigfield elements. This means that a call to the same field element twice to...
According the description of the algorithm, this shouldn't happen. Investigate
When we retrieve monomials in IPA or KZG they don't check that there are enough points. Then we get bugs down the line while trying to access them,
Since we are not going to be doing degree checks on translator polynomials in the mini circuit, we have to add relations ensuring range constraint wires are zeroed out outside...
Comparing to literal 0 costs 35 ns, which is too much just for a comparison, because it is inefficient.
A mechanism that treats the whole circuit before finalization (in case of ultra) as a graph with additional information and looks at the following properties: 1) How many independent subgraphs...
Processing RAM/ROM updates range constraints during finalization, so computation might be off
Currently gemini breaks for a polynomial, where the second half of coefficients is zero.
We need to ensure no parts of the protocol use transcript in an unexpected way. We can do tracking of usage of particular values with DFSan