Merge pull request #636 from mir-protocol/challenger_fixed_buffer

Use a fixed input buffer size in `Challenger`.
This commit is contained in:
Daniel Lubarov 2022-07-26 16:09:19 -07:00 committed by GitHub
commit 6df1a669e1
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

View File

@ -33,8 +33,8 @@ impl<F: RichField, H: Hasher<F>> Challenger<F, H> {
pub fn new() -> Challenger<F, H> {
Challenger {
sponge_state: [F::ZERO; SPONGE_WIDTH],
input_buffer: Vec::new(),
output_buffer: Vec::new(),
input_buffer: Vec::with_capacity(SPONGE_RATE),
output_buffer: Vec::with_capacity(SPONGE_RATE),
_phantom: Default::default(),
}
}
@ -44,6 +44,10 @@ impl<F: RichField, H: Hasher<F>> Challenger<F, H> {
self.output_buffer.clear();
self.input_buffer.push(element);
if self.input_buffer.len() == SPONGE_RATE {
self.duplexing();
}
}
pub fn observe_extension_element<const D: usize>(&mut self, element: &F::Extension)
@ -79,12 +83,10 @@ impl<F: RichField, H: Hasher<F>> Challenger<F, H> {
}
pub fn get_challenge(&mut self) -> F {
self.absorb_buffered_inputs();
if self.output_buffer.is_empty() {
// Evaluate the permutation to produce `r` new outputs.
self.sponge_state = H::Permutation::permute(self.sponge_state);
self.output_buffer = self.sponge_state[0..SPONGE_RATE].to_vec();
// If we have buffered inputs, we must perform a duplexing so that the challenge will
// reflect them. Or if we've run out of outputs, we must perform a duplexing to get more.
if !self.input_buffer.is_empty() || self.output_buffer.is_empty() {
self.duplexing();
}
self.output_buffer
@ -125,27 +127,24 @@ impl<F: RichField, H: Hasher<F>> Challenger<F, H> {
.collect()
}
/// Absorb any buffered inputs. After calling this, the input buffer will be empty.
fn absorb_buffered_inputs(&mut self) {
if self.input_buffer.is_empty() {
return;
/// Absorb any buffered inputs. After calling this, the input buffer will be empty, and the
/// output buffer will be full.
fn duplexing(&mut self) {
assert!(self.input_buffer.len() <= SPONGE_RATE);
// Overwrite the first r elements with the inputs. This differs from a standard sponge,
// where we would xor or add in the inputs. This is a well-known variant, though,
// sometimes called "overwrite mode".
for (i, input) in self.input_buffer.drain(..).enumerate() {
self.sponge_state[i] = input;
}
for input_chunk in self.input_buffer.chunks(SPONGE_RATE) {
// Overwrite the first r elements with the inputs. This differs from a standard sponge,
// where we would xor or add in the inputs. This is a well-known variant, though,
// sometimes called "overwrite mode".
for (i, &input) in input_chunk.iter().enumerate() {
self.sponge_state[i] = input;
}
// Apply the permutation.
self.sponge_state = H::Permutation::permute(self.sponge_state);
// Apply the permutation.
self.sponge_state = H::Permutation::permute(self.sponge_state);
}
self.output_buffer = self.sponge_state[0..SPONGE_RATE].to_vec();
self.input_buffer.clear();
self.output_buffer.clear();
self.output_buffer
.extend_from_slice(&self.sponge_state[0..SPONGE_RATE]);
}
}
@ -155,7 +154,9 @@ impl<F: RichField, H: AlgebraicHasher<F>> Default for Challenger<F, H> {
}
}
/// A recursive version of `Challenger`.
/// A recursive version of `Challenger`. The main difference is that `RecursiveChallenger`'s input
/// buffer can grow beyond `SPONGE_RATE`. This is so that `observe_element` etc do not need access
/// to the `CircuitBuilder`.
pub struct RecursiveChallenger<F: RichField + Extendable<D>, H: AlgebraicHasher<F>, const D: usize>
{
sponge_state: [Target; SPONGE_WIDTH],
@ -248,7 +249,8 @@ impl<F: RichField + Extendable<D>, H: AlgebraicHasher<F>, const D: usize>
self.get_n_challenges(builder, D).try_into().unwrap()
}
/// Absorb any buffered inputs. After calling this, the input buffer will be empty.
/// Absorb any buffered inputs. After calling this, the input buffer will be empty, and the
/// output buffer will be full.
fn absorb_buffered_inputs(&mut self, builder: &mut CircuitBuilder<F, D>) {
if self.input_buffer.is_empty() {
return;