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
Currently, the inner-product prover splits vectors using pairs of slices. In each case, the intent is to split each vector in half for folding operations.
A somewhat more idiomatic approach could be to perform a split directly.
This could also be a good opportunity to perform bounds checking in each inner-product round, instead of only initially, just in case the caller did something wonky between rounds.
The text was updated successfully, but these errors were encountered:
The inner-product prover is a bit clunky, as it requires the range
prover to set up a struct, perform iterative rounds, and then parse out
the proof elements it needs. This also requires consistency checks that
it would be nicer to avoid.
This PR moves the inner-product prover into the range prover, which
simplifies things at the cost of a longer prover function.
It also fixes some [vector
allocations](https://github.com/tari-project/bulletproofs-plus/blob/d9d0cc9063f85684179908569227dda251981751/src/inner_product_round.rs#L111-L112)
that were much bigger than necessary.
Closes#61. Closes#55. Supersedes #56.
Currently, the inner-product prover splits vectors using pairs of slices. In each case, the intent is to split each vector in half for folding operations.
A somewhat more idiomatic approach could be to perform a split directly.
This could also be a good opportunity to perform bounds checking in each inner-product round, instead of only initially, just in case the caller did something wonky between rounds.
The text was updated successfully, but these errors were encountered: