Test vectors & benchmark for PCS (argumentcomputer#222)

* test(ipa): proof verified random seed

test(ipa): proof verified

test(ipa): switch from keccak to grumpkin transcript

test(ipa): wip test IPA prove & verify

chore: added Jetbrain config to gitignore

* test(ipa): using generic fn over mlkzg, ipa & zm

- Implemented test_fail_bad_proof that generates a test case from a seed to try and verify a non valid proof.
- Implement both test_fail_bad_proof & test_from_seed for IPA over Grumpkin, MLKZG over Bn256 and ZM over Bn256

* ci(ipa): fixed xclippy

* feat(ipa): wip bench pcs

* feat(ipa): bench proof generation and verification

- Created a benchmark for proof generation and verification for IPA over Grumpkin, MLKZG over Bn and ZM over Bn
- multilinear.rs to public for benchmark purposes
- Refactored generic test methods per @adr1anh comment

* feat(ipa): proper polynomial sizes for test & sample size to 10

* feat(ipa): benchmark groups

* ci(ipa): fix xclippy

* test(ipa): added more polynomial sizes

* test(ipa): renamed test method

* refactor(ipa): revamp based on review

* refactor(ipa): Flat -> Auto bench

* refactor(ipa): random_with_eval inline

* ci(ipa): fix xclippy

.gitignore

@@ -10,4 +10,7 @@ Cargo.lock
 **/*.rs.bk
 
 # Configurations for VSCode
-.vscode/
+.vscode/
+
+# Configuration for Jetbrains
+.idea/

Cargo.toml

@@ -96,6 +96,10 @@ harness = false
 name = "compressed-snark-supernova"
 harness = false
 
+[[bench]]
+name = "pcs"
+harness = false
+
 [features]
 default = []
 abomonate = []

benches/compressed-snark.rs

@@ -27,7 +27,7 @@ type SS2 = arecibo::spartan::ppsnark::RelaxedR1CSSNARK<E2, EE2>;
 type C1 = NonTrivialCircuit<<E1 as Engine>::Scalar>;
 type C2 = TrivialCircuit<<E2 as Engine>::Scalar>;
 
-// To run these benchmarks, first download `criterion` with `cargo install cargo install cargo-criterion`.
+// To run these benchmarks, first download `criterion` with `cargo install cargo-criterion`.
 // Then `cargo criterion --bench compressed-snark`. The results are located in `target/criterion/data/<name-of-benchmark>`.
 // For flamegraphs, run `cargo criterion --bench compressed-snark --features flamegraph -- --profile-time <secs>`.
 // The results are located in `target/criterion/profile/<name-of-benchmark>`.

benches/pcs.rs

@@ -0,0 +1,198 @@
+use arecibo::provider::{
+  ipa_pc::EvaluationEngine as IPAEvaluationEngine, mlkzg::EvaluationEngine as MLEvaluationEngine,
+  non_hiding_zeromorph::ZMPCS, Bn256EngineKZG, Bn256EngineZM, GrumpkinEngine,
+};
+use arecibo::spartan::polys::multilinear::MultilinearPolynomial;
+use arecibo::traits::{
+  commitment::CommitmentEngineTrait, evaluation::EvaluationEngineTrait, Engine,
+  TranscriptEngineTrait,
+};
+use criterion::{criterion_group, criterion_main, Bencher, BenchmarkId, Criterion, SamplingMode};
+use ff::Field;
+use halo2curves::bn256::Bn256;
+use rand::rngs::StdRng;
+use rand_core::{CryptoRng, RngCore, SeedableRng};
+use std::any::type_name;
+use std::time::Duration;
+
+// To run these benchmarks, first download `criterion` with `cargo install cargo-criterion`.
+// Then `cargo criterion --bench pcs`.
+// For flamegraphs, run `cargo criterion --bench pcs --features flamegraph -- --profile-time <secs>`.
+// The results are located in `target/criterion/profile/<name-of-benchmark>`.
+cfg_if::cfg_if! {
+  if #[cfg(feature = "flamegraph")] {
+    criterion_group! {
+          name = pcs;
+          config = Criterion::default().warm_up_time(Duration::from_millis(3000)).with_profiler(pprof::criterion::PProfProfiler::new(100, pprof::criterion::Output::Flamegraph(None)));
+          targets = bench_pcs
+    }
+  } else {
+    criterion_group! {
+          name = pcs;
+          config = Criterion::default().warm_up_time(Duration::from_millis(3000));
+          targets = bench_pcs
+    }
+  }
+}
+
+criterion_main!(pcs);
+
+const NUM_VARS_TEST_VECTOR: [usize; 6] = [10, 12, 14, 16, 18, 20];
+
+struct BenchAssests<E: Engine, EE: EvaluationEngineTrait<E>> {
+  poly: MultilinearPolynomial<<E as Engine>::Scalar>,
+  point: Vec<<E as Engine>::Scalar>,
+  eval: <E as Engine>::Scalar,
+  ck: <<E as Engine>::CE as CommitmentEngineTrait<E>>::CommitmentKey,
+  commitment: <<E as Engine>::CE as CommitmentEngineTrait<E>>::Commitment,
+  prover_key: <EE as EvaluationEngineTrait<E>>::ProverKey,
+  verifier_key: <EE as EvaluationEngineTrait<E>>::VerifierKey,
+  proof: Option<<EE as EvaluationEngineTrait<E>>::EvaluationArgument>,
+}
+
+/// Returns a random polynomial, a point and calculate its evaluation.
+pub fn random_poly_with_eval<E: Engine, R: RngCore + CryptoRng>(
+  num_vars: usize,
+  mut rng: &mut R,
+) -> (
+  MultilinearPolynomial<<E as Engine>::Scalar>,
+  Vec<<E as Engine>::Scalar>,
+  <E as Engine>::Scalar,
+) {
+  // Generate random polynomial and point.
+  let poly = MultilinearPolynomial::random(num_vars, &mut rng);
+  let point = (0..num_vars)
+    .map(|_| <E as Engine>::Scalar::random(&mut rng))
+    .collect::<Vec<_>>();
+
+  // Calculation evaluation of point over polynomial.
+  let eval = MultilinearPolynomial::evaluate_with(poly.evaluations(), &point);
+
+  (poly, point, eval)
+}
+
+impl<E: Engine, EE: EvaluationEngineTrait<E>> BenchAssests<E, EE> {
+  pub(crate) fn from_num_vars<R: CryptoRng + RngCore>(num_vars: usize, rng: &mut R) -> Self {
+    let (poly, point, eval) = random_poly_with_eval::<E, R>(num_vars, rng);
+
+    // Mock commitment key.
+    let ck = E::CE::setup(b"test", 1 << num_vars);
+    // Commits to the provided vector using the provided generators.
+    let commitment = E::CE::commit(&ck, poly.evaluations());
+
+    let (prover_key, verifier_key) = EE::setup(&ck);
+
+    // Generate proof so that we can bench verification.
+    let proof = EE::prove(
+      &ck,
+      &prover_key,
+      &mut E::TE::new(b"TestEval"),
+      &commitment,
+      poly.evaluations(),
+      &point,
+      &eval,
+    )
+    .unwrap();
+
+    Self {
+      poly,
+      point,
+      eval,
+      ck,
+      commitment,
+      prover_key,
+      verifier_key,
+      proof: Some(proof),
+    }
+  }
+}
+
+// Macro to generate benchmark code for multiple evaluation engine types
+macro_rules! benchmark_all_engines {
+    ($criterion:expr, $test_vector:expr, $proving_fn:expr, $verifying_fn:expr, $( ($assets:ident, $eval_engine:ty) ),*) => {
+        for num_vars in $test_vector.iter() {
+            let mut rng = rand::rngs::StdRng::seed_from_u64(*num_vars as u64);
+
+            $(
+                let $assets: BenchAssests<_, $eval_engine> = BenchAssests::from_num_vars::<StdRng>(*num_vars, &mut rng);
+            )*
+
+            // Proving group
+            let mut proving_group = $criterion.benchmark_group(format!("PCS-Proving {}", num_vars));
+            proving_group
+                .sampling_mode(SamplingMode::Auto)
+                .sample_size(10);
+
+            $(
+                proving_group.bench_with_input(BenchmarkId::new(type_name::<$eval_engine>(), num_vars), &num_vars, |b, _| {
+                    $proving_fn(b, &$assets);
+                });
+            )*
+
+            proving_group.finish();
+
+            // Verifying group
+            let mut verifying_group = $criterion.benchmark_group(format!("PCS-Verifying {}", num_vars));
+            verifying_group
+                .sampling_mode(SamplingMode::Auto)
+                .sample_size(10);
+
+            $(
+                verifying_group.bench_with_input(BenchmarkId::new(type_name::<$eval_engine>(), num_vars), &num_vars, |b, _| {
+                    $verifying_fn(b, &$assets);
+                });
+            )*
+
+            verifying_group.finish();
+        }
+    };
+}
+
+fn bench_pcs(c: &mut Criterion) {
+  benchmark_all_engines!(
+    c,
+    NUM_VARS_TEST_VECTOR,
+    bench_pcs_proving_internal,
+    bench_pcs_verifying_internal,
+    (ipa_assets, IPAEvaluationEngine<GrumpkinEngine>),
+    (mlkzg_assets, MLEvaluationEngine<Bn256, Bn256EngineKZG>),
+    (zm_assets, ZMPCS<Bn256, Bn256EngineZM>)
+  );
+}
+
+fn bench_pcs_proving_internal<E: Engine, EE: EvaluationEngineTrait<E>>(
+  b: &mut Bencher<'_>,
+  bench_assets: &BenchAssests<E, EE>,
+) {
+  // Bench generate proof.
+  b.iter(|| {
+    EE::prove(
+      &bench_assets.ck,
+      &bench_assets.prover_key,
+      &mut E::TE::new(b"TestEval"),
+      &bench_assets.commitment,
+      bench_assets.poly.evaluations(),
+      &bench_assets.point,
+      &bench_assets.eval,
+    )
+    .unwrap();
+  });
+}
+
+fn bench_pcs_verifying_internal<E: Engine, EE: EvaluationEngineTrait<E>>(
+  b: &mut Bencher<'_>,
+  bench_assets: &BenchAssests<E, EE>,
+) {
+  // Bench verify proof.
+  b.iter(|| {
+    EE::verify(
+      &bench_assets.verifier_key,
+      &mut E::TE::new(b"TestEval"),
+      &bench_assets.commitment,
+      &bench_assets.point,
+      &bench_assets.eval,
+      bench_assets.proof.as_ref().unwrap(),
+    )
+    .unwrap();
+  });
+}

benches/recursive-snark.rs

@@ -19,7 +19,7 @@ type E2 = VestaEngine;
 type C1 = NonTrivialCircuit<<E1 as Engine>::Scalar>;
 type C2 = TrivialCircuit<<E2 as Engine>::Scalar>;
 
-// To run these benchmarks, first download `criterion` with `cargo install cargo install cargo-criterion`.
+// To run these benchmarks, first download `criterion` with `cargo install cargo-criterion`.
 // Then `cargo criterion --bench recursive-snark`. The results are located in `target/criterion/data/<name-of-benchmark>`.
 // For flamegraphs, run `cargo criterion --bench recursive-snark --features flamegraph -- --profile-time <secs>`.
 // The results are located in `target/criterion/profile/<name-of-benchmark>`.

src/provider/ipa_pc.rs

@@ -410,3 +410,17 @@ where
     }
   }
 }
+
+#[cfg(test)]
+mod test {
+  use crate::provider::ipa_pc::EvaluationEngine;
+  use crate::provider::util::test_utils::prove_verify_from_num_vars;
+  use crate::provider::GrumpkinEngine;
+
+  #[test]
+  fn test_multiple_polynomial_size() {
+    for num_vars in [4, 5, 6] {
+      prove_verify_from_num_vars::<_, EvaluationEngine<GrumpkinEngine>>(num_vars);
+    }
+  }
+}

src/provider/mlkzg.rs

@@ -429,12 +429,11 @@ where
 #[cfg(test)]
 mod tests {
   use super::*;
+  use crate::provider::util::test_utils::prove_verify_from_num_vars;
   use crate::{
-    provider::keccak::Keccak256Transcript, spartan::polys::multilinear::MultilinearPolynomial,
-    traits::commitment::CommitmentTrait, CommitmentKey,
+    provider::keccak::Keccak256Transcript, traits::commitment::CommitmentTrait, CommitmentKey,
   };
   use bincode::Options;
-  use rand::SeedableRng;
 
   type E = halo2curves::bn256::Bn256;
   type NE = crate::provider::Bn256EngineKZG;
@@ -558,54 +557,8 @@ mod tests {
   #[test]
   fn test_mlkzg_more() {
     // test the mlkzg prover and verifier with random instances (derived from a seed)
-    for ell in [4, 5, 6] {
-      let mut rng = rand::rngs::StdRng::seed_from_u64(ell as u64);
-
-      let n = 1 << ell; // n = 2^ell
-
-      let poly = (0..n).map(|_| Fr::random(&mut rng)).collect::<Vec<_>>();
-      let point = (0..ell).map(|_| Fr::random(&mut rng)).collect::<Vec<_>>();
-      let eval = MultilinearPolynomial::evaluate_with(&poly, &point);
-
-      let ck: CommitmentKey<NE> =
-        <KZGCommitmentEngine<E> as CommitmentEngineTrait<NE>>::setup(b"test", n);
-      let (pk, vk): (KZGProverKey<E>, KZGVerifierKey<E>) = EvaluationEngine::<E, NE>::setup(&ck);
-
-      // make a commitment
-      let C = <KZGCommitmentEngine<E> as CommitmentEngineTrait<NE>>::commit(&ck, &poly);
-
-      // prove an evaluation
-      let mut prover_transcript = Keccak256Transcript::<NE>::new(b"TestEval");
-      let proof: EvaluationArgument<E> = EvaluationEngine::<E, NE>::prove(
-        &ck,
-        &pk,
-        &mut prover_transcript,
-        &C,
-        &poly,
-        &point,
-        &eval,
-      )
-      .unwrap();
-
-      // verify the evaluation
-      let mut verifier_tr = Keccak256Transcript::<NE>::new(b"TestEval");
-      assert!(
-        EvaluationEngine::<E, NE>::verify(&vk, &mut verifier_tr, &C, &point, &eval, &proof).is_ok()
-      );
-
-      // Change the proof and expect verification to fail
-      let mut bad_proof = proof.clone();
-      bad_proof.comms[0] = (bad_proof.comms[0] + bad_proof.comms[1]).to_affine();
-      let mut verifier_tr2 = Keccak256Transcript::<NE>::new(b"TestEval");
-      assert!(EvaluationEngine::<E, NE>::verify(
-        &vk,
-        &mut verifier_tr2,
-        &C,
-        &point,
-        &eval,
-        &bad_proof
-      )
-      .is_err());
+    for num_vars in [4, 5, 6] {
+      prove_verify_from_num_vars::<_, EvaluationEngine<E, NE>>(num_vars);
     }
   }
 }

src/provider/mod.rs

@@ -15,7 +15,7 @@ pub(crate) mod traits;
 // a non-hiding variant of {kzg, zeromorph}
 pub(crate) mod kzg_commitment;
 pub(crate) mod non_hiding_kzg;
-mod util;
+pub(crate) mod util;
 
 // crate-private modules
 mod keccak;