Marc/lookup prover #3008
Marc/lookup prover #3008
Conversation
marcbeunardeau88
force-pushed
the
marc/lookup-prover
branch
from
4a9c7bf to
cadf895
Compare
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #3008 +/- ##
==========================================
- Coverage 77.15% 76.46% -0.69%
==========================================
Files 260 263 +3
Lines 61529 62170 +641
==========================================
+ Hits 47470 47540 +70
- Misses 14059 14630 +571 ☔ View full report in Codecov by Sentry. |
| use o1_utils::ExtendedDensePolynomial; | ||
| use poly_commitment::{commitment::absorb_commitment, ipa::SRS, OpenProof, SRS as _}; | ||
| //TODO Parralelize | ||
| //use rayon::prelude::*; |
| } | ||
| // TODO: I could not find a more elegant solution to map over this struct | ||
| impl<X> ColumnEnv<X> { | ||
| pub fn my_map<Y, F>(self, f: F) -> ColumnEnv<Y> |
There was a problem hiding this comment.
It feels like this would be simpler as just a normal map implementation, like
ColumnEnv {
wires: self.wires.into_iter().map(f).collect(),
...
}and the known iterator sizes will help the vector allocator to allocate the right amount of memory instead of growing to meet it.
There was a problem hiding this comment.
Also, map is probably preferable to my_map as a name.
| #[derive(Clone)] | ||
| pub struct AllColumns<X> { | ||
| pub cols: ColumnEnv<X>, | ||
| pub t_shares: Vec<X>, |
There was a problem hiding this comment.
What is a t_shares? This would benefit from a descriptive name
| type IntoIter = | ||
| Chain<<ColumnEnv<X> as IntoIterator>::IntoIter, <Vec<X> as IntoIterator>::IntoIter>; | ||
| fn into_iter(self) -> Self::IntoIter { | ||
| self.cols.into_iter().chain(self.t_shares) |
There was a problem hiding this comment.
Style nit: here and elsewhere, using a destructuring assignment for self will give you a warning if new fields are ever added
adds a prover for a delayed lookup argument
Proves that we can accumulate some lookup, so that the sum goes from a public input acc_init to a public output acc_final.
A lot of unecessarry cloning is happening, which will be handle in a follow-up.
A lot of vec are used instead of array, which will be handle in a foloow-up.