Upgrade to v0.4.0 by TakodaS · Pull Request #94 · arkworks-rs/marlin

TakodaS · 2023-09-07T14:41:25Z

Description

Updated arkworks-rs dependencies to v0.4.0.

Before we can merge this PR, please make sure that all the following items have been
checked off. If any of the checklist items are not applicable, please leave them but
write a little note why.

Targeted PR against correct branch (master)
Linked to Github issue with discussion and accepted design OR have an explanation in the PR that describes this work.
Wrote unit tests (Not needed as no new functionality was added).
Updated relevant documentation in the code
Added a relevant changelog entry to the Pending section in CHANGELOG.md
Re-reviewed Files changed in the Github PR explorer

TakodaS · 2023-09-07T14:43:41Z

@Pratyush Should I bother implementing CryptographicSponge for the Fiat-Shamir hash function found in rng or should I rip this all out in favour of Poseidon?

Edit: I see something similar is already in PR. Maybe I can commandeer that PR. I will go talk to @vlopes11.

TakodaS · 2023-09-12T16:26:10Z

@Pratyush Should I bother implementing CryptographicSponge for the Fiat-Shamir hash function found in rng or should I rip this all out in favour of Poseidon?

Edit: I see something similar is already in PR. Maybe I can commandeer that PR. I will go talk to @vlopes11.

Implementing FS hashing using the PoseidonSponge, temporarily with the same default initialization as used in the tests for crypto-primitives.

TakodaS · 2023-09-28T11:03:03Z

Cargo.toml

 rayon = { version = "1", optional = true }
 digest = { version = "0.9" }
 derivative = { version = "2", features = ["use_core"] }
+itertools = "0.11.0"


I will probably have to remove this to pass the no-std test.

TakodaS · 2023-09-28T11:04:50Z

src/rng.rs

+#[derive(Clone)]
+pub struct SimplePoseidonRng<F: PrimeField>(PoseidonSponge<F>);
+
+impl<F: PrimeField> RngCore for SimplePoseidonRng<F> {


RngCore is required for backwards compatibility with generation of random field elements in ark-ff and ark-poly. Changing these is a much larger piece of work than exposing the sponge construction through RngCore and beyond the scope of this PR imo.

TakodaS · 2023-09-28T11:06:58Z

src/rng.rs

+/// Instantiate Poseidon sponge with default parameters
+impl<F: PrimeField> Default for SimplePoseidonRng<F> {
+    fn default() -> Self {
+        // let default =
+        // Self(PoseidonSponge::new(&poseidon_parameters_for_test()))
+        let (alpha, rate, full_rounds, partial_rounds) = (17, 2, 8, 29);
+        let (ark, mds) = find_poseidon_ark_and_mds(
+            F::MODULUS_BIT_SIZE as u64,
+            rate,
+            full_rounds,
+            partial_rounds,
+            0,
+        );
+        let config = PoseidonConfig {
+            full_rounds: full_rounds as usize,
+            partial_rounds: partial_rounds as usize,
+            alpha: alpha as u64,
+            ark,
+            mds,
+            rate,
+            capacity: 1,
+        };
+        SimplePoseidonRng(PoseidonSponge::new(&config))


I hope these are good defaults.

TakodaS · 2023-10-03T14:55:35Z

src/lib.rs

-impl<F: PrimeField, PC: PolynomialCommitment<F, DensePolynomial<F>>, FS: FiatShamirRng>
-    Marlin<F, PC, FS>
+impl<
+        F: PrimeField + Absorb,


The addition of the Absorb trait bound is very problematic. Although it will improve efficiency of absorption, it makes the implementation far less generic. I am not certain how to improve this, considering that specialization of traits is an unstable feature.

I resolved this by implementing the fast_prove and fast_verify methods that may be invoked when the PrimeField also satisfies the Absorb trait bound. I think this is a good trade off, since Absorb is not nearly as well implemented as Canonical(De)Serialize and it may be a pain for people to upgrade to the new implementation. Over time, I think the old method can be deprecated in favor of the new, but I think a prerequisite is implementing a derive Absorb macro.

Obviously there is a lot of code duplication, this can be refactored if @Pratyush approves the design philosophy here.

autquis · 2024-02-13T23:54:20Z

Any updates on this? I am ready to take any necessary steps to make it ready to merge. @Pratyush @mmagician

Pratyush · 2024-02-27T22:11:00Z

src/ahp/constraint_systems.rs

+        let (eta_a, eta_b, eta_c) = rng
+            .squeeze_field_elements(3)
+            .iter()
+            .map(|x: &F| x.to_owned())
+            .collect_tuple()
+            .unwrap();


Suggested change

let (eta_a, eta_b, eta_c) = rng

.squeeze_field_elements(3)

.iter()

.map(|x: &F| x.to_owned())

.collect_tuple()

.unwrap();

let [eta_a, eta_b, eta_c] = rng

.squeeze_field_elements(3)[..3]

else { unreachable!("should have three elements") };

Pratyush · 2024-02-27T22:13:37Z

src/ahp/prover.rs

+        let (f1, f2, f3) = rng
+            .squeeze_field_elements(3)
+            .iter()
+            .map(|x: &F| x.to_owned())
+            .collect_tuple()
+            .unwrap();


This is not secure; we cannot use the sponge for randomness for zero knowledge. The old version should work fine.

Pratyush · 2024-02-27T22:14:42Z

src/ahp/verifier.rs

+        let (eta_a, eta_b, eta_c) = rng
+            .squeeze_field_elements(3)
+            .iter()
+            .map(|x: &F| x.to_owned())
+            .collect_tuple()
+            .unwrap();


Suggested change

let (eta_a, eta_b, eta_c) = rng

.squeeze_field_elements(3)

.iter()

.map(|x: &F| x.to_owned())

.collect_tuple()

.unwrap();

let [eta_a, eta_b, eta_c, ..] = rng

.squeeze_field_elements(3)[..3]

else { unreachable!("should be of size 3) };

Pratyush · 2024-02-27T22:16:55Z

src/lib.rs

+        let fcinput = first_comms
+            .iter()
+            .map(|p| p.commitment().clone())
+            .collect::<Vec<_>>();

-        fs_rng.absorb(&to_bytes![first_comms, prover_first_msg].unwrap());
+        match prover_first_msg {
+            ProverMsg::FieldElements(ref elems) => {
+                absorb!(&mut fs_rng, &to_bytes(&fcinput), &to_bytes(elems));
+            }
+            ProverMsg::EmptyMessage => fs_rng.absorb(&to_bytes(&fcinput)),
+        }


Let's wrap this into a helper function.

Pratyush · 2024-02-27T22:17:02Z

src/lib.rs

+        let scinput = second_comms
+            .iter()
+            .map(|p| p.commitment().clone())
+            .collect::<Vec<_>>();
+        match prover_second_msg {
+            ProverMsg::FieldElements(ref elems) => {
+                absorb!(&mut fs_rng, &to_bytes(&scinput), &to_bytes(elems));
+            }
+            ProverMsg::EmptyMessage => fs_rng.absorb(&to_bytes(&scinput)),
+        }


Pratyush · 2024-02-27T22:17:08Z

src/lib.rs

+        let tcinput = third_comms
+            .iter()
+            .map(|p| p.commitment().clone())
+            .collect::<Vec<_>>();
+        match prover_third_msg {
+            ProverMsg::FieldElements(ref elems) => {
+                absorb!(&mut fs_rng, &to_bytes(&tcinput), &to_bytes(elems));
+            }
+            ProverMsg::EmptyMessage => fs_rng.absorb(&to_bytes(&tcinput)),
+        }


Pratyush · 2024-02-27T22:17:46Z

src/lib.rs

+        match &proof.prover_messages[0] {
+            ProverMsg::FieldElements(ref elems) => {
+                absorb!(&mut fs_rng, &to_bytes(first_comms), &to_bytes(elems));
+            }
+            ProverMsg::EmptyMessage => fs_rng.absorb(&to_bytes(first_comms)),
+        }
+        let (_, verifier_state) =
+            AHPForR1CS::verifier_first_round(index_vk.index_info, &mut fs_rng)?;


Let's use the helper function here also.

Pratyush · 2024-02-27T22:19:27Z

src/lib.rs

+{
+    /// Create a zkSNARK asserting that the constraint system is satisfied.
+    /// Uses fast absorption of field elements into sponge
+    pub fn fast_prove<C: ConstraintSynthesizer<F>, R: RngCore + CryptographicSponge>(


What is this new function?

#94 (comment)

Pratyush · 2024-02-27T22:20:14Z

src/test.rs

+            let (a, b) = rng
+                .squeeze_field_elements(2)
+                .iter()
+                .map(|x: &Fr| x.to_owned())
+                .collect_tuple()
+                .unwrap();


Let's create a helper squeeze macro.

TakodaS · 2024-03-06T10:25:12Z

Any updates on this? I am ready to take any necessary steps to make it ready to merge. @Pratyush @mmagician

Go right ahead! I guess you can make a new pull request and I will close this one?

TakodaS added 2 commits September 7, 2023 14:34

updated deps

2e93ba9

implemented canonical serialization

a4b81ef

TakodaS marked this pull request as draft September 7, 2023 14:41

TakodaS marked this pull request as ready for review September 7, 2023 16:15

TakodaS changed the title ~~Draft: Upgrade to v0.4.0~~ Upgrade to v0.4.0 Sep 7, 2023

TakodaS added 11 commits September 8, 2023 12:22

moved sponge and fiat shamir in

c7ae172

changed dependencies to ark_r1cs lib and fixes

4d616f4

checkin

5d6d7d0

serialize mostly done

aa666b5

readd old rng mod

9bdd8a2

remove pedanticerrors temporarily

71a603b

updated realtive to poly-commit-v0.4

c3dbc05

implemented poseidon as rng

cbf5a4a

added default and spong methods

d860a91

LabelledCommitment does not implement CanonicalSerialize

66a6786

builds!

9ab0b3e

TakodaS mentioned this pull request Sep 12, 2023

Add serialize traits to LabeledCommitment arkworks-rs/poly-commit#122

Closed

6 tasks

TakodaS added 8 commits September 18, 2023 12:40

tests building

e85c8fc

to unlabeled commit

d805506

outlinign test pass

1df70b3

revert to ark poly commit

8a91943

bench building

d1b8e51

fmt and fix

b5071fc

imports

fed7d03

unmodified .gitignore

0c57285

TakodaS force-pushed the master branch from ef3e575 to 0c57285 Compare September 19, 2023 11:45

TakodaS added 2 commits September 19, 2023 13:07

updated changelog

f07a64d

removed unnecessary r1cs-std dependency

8b55468

TakodaS added 13 commits September 20, 2023 16:19

easy changes

f03a658

refactor RNG initialization

127dc2f

before test

6f11b89

removed some hardcoded poseidon config parameters

2a70370

semi implemented sponge methods

a2c56bf

native absorption of field elements

cfefa0f

tests passing

f2152b3

benches running

6b18bba

removed to_bytes!

a3d9ba6

cargo fix+fmt

3f886a9

removed unnecessary trait

11bd766

reintroduce optimization of tests after debugging

c5ab30b

update changelog

788a34e

TakodaS commented Sep 28, 2023

View reviewed changes

TakodaS requested a review from Pratyush September 28, 2023 11:05

TakodaS commented Sep 28, 2023

View reviewed changes

TakodaS commented Oct 3, 2023

View reviewed changes

TakodaS added 2 commits October 9, 2023 15:52

added fast_prove and verify methods

093fdcc

added test of fast_proof and fast_verify

6a071f7

Pratyush reviewed Feb 27, 2024

View reviewed changes

Conversation

TakodaS commented Sep 7, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Uh oh!

TakodaS commented Sep 7, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

TakodaS commented Sep 12, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

TakodaS Oct 9, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

autquis commented Feb 13, 2024

Uh oh!

Pratyush Feb 27, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

TakodaS commented Mar 6, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

TakodaS commented Sep 7, 2023 •

edited

Loading

TakodaS commented Sep 7, 2023 •

edited

Loading

TakodaS commented Sep 12, 2023 •

edited

Loading

TakodaS Oct 9, 2023 •

edited

Loading

Pratyush Feb 27, 2024 •

edited

Loading

TakodaS commented Mar 6, 2024 •

edited

Loading