Update toolchain

.github/workflows/benchmarks-pages.yaml

@@ -1,4 +1,4 @@
-name: 
+name:
 
 on:
   push:
@@ -18,7 +18,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - name: Run benchmark
         run: ./scripts/bench.sh -- --output-format bencher | tee output.txt
       - name: Download previous benchmark data
@@ -29,7 +29,7 @@ jobs:
       - name: Store benchmark result
         uses: benchmark-action/github-action-benchmark@v1
         with:
-          tool: 'cargo'
+          tool: "cargo"
           output-file-path: output.txt
           github-token: ${{ secrets.GITHUB_TOKEN }}
           auto-push: true

.github/workflows/ci.yaml

@@ -25,7 +25,7 @@ jobs:
       - uses: dtolnay/rust-toolchain@master
         with:
           components: rustfmt
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - uses: Swatinem/rust-cache@v2
       - run: scripts/rust_fmt.sh --check
 
@@ -36,7 +36,7 @@ jobs:
       - uses: dtolnay/rust-toolchain@master
         with:
           components: clippy
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - uses: Swatinem/rust-cache@v2
       - run: scripts/clippy.sh
 
@@ -46,17 +46,17 @@ jobs:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - uses: Swatinem/rust-cache@v2
-      - run: cargo +nightly-2024-01-04 doc
+      - run: cargo +nightly-2024-09-21 doc
 
   run-wasm32-wasi-tests:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
           targets: wasm32-wasi
       - uses: taiki-e/install-action@v2
         with:
@@ -73,7 +73,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
           targets: wasm32-unknown-unknown
       - uses: Swatinem/rust-cache@v2
       - uses: jetli/[email protected]
@@ -89,9 +89,9 @@ jobs:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - uses: Swatinem/rust-cache@v2
-      - run: cargo +nightly-2024-01-04 test
+      - run: cargo +nightly-2024-09-21 test
         env:
           RUSTFLAGS: -C target-feature=+neon
 
@@ -104,9 +104,9 @@ jobs:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - uses: Swatinem/rust-cache@v2
-      - run: cargo +nightly-2024-01-04 test
+      - run: cargo +nightly-2024-09-21 test
         env:
           RUSTFLAGS: -C target-cpu=native -C target-feature=+${{ matrix.target-feature }}
 
@@ -116,7 +116,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - name: Run benchmark
         run: ./scripts/bench.sh -- --output-format bencher | tee output.txt
       - name: Download previous benchmark data
@@ -142,19 +142,19 @@ jobs:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - uses: Swatinem/rust-cache@v2
-      - run: cargo +nightly-2024-01-04 test
+      - run: cargo +nightly-2024-09-21 test
 
   run-slow-tests:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - uses: Swatinem/rust-cache@v2
-      - run: cargo +nightly-2024-01-04 test --release --features="slow-tests"
+      - run: cargo +nightly-2024-09-21 test --release --features="slow-tests"
 
   udeps:
     runs-on: ubuntu-latest
@@ -163,7 +163,7 @@ jobs:
       - uses: dtolnay/rust-toolchain@master
         name: "Rust Toolchain Setup"
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - uses: Swatinem/rust-cache@v2
         id: "cache-cargo"
       - if: ${{ steps.cache-cargo.outputs.cache-hit != 'true' }}
@@ -172,7 +172,7 @@ jobs:
           wget -O - -c https://github.com/est31/cargo-udeps/releases/download/v0.1.35/cargo-udeps-v0.1.35-x86_64-unknown-linux-gnu.tar.gz | tar -xz
           cargo-udeps-*/cargo-udeps udeps
         env:
-          RUSTUP_TOOLCHAIN: nightly-2024-01-04
+          RUSTUP_TOOLCHAIN: nightly-2024-09-21
 
   all-tests:
     runs-on: ubuntu-latest

.github/workflows/coverage.yaml

@@ -12,14 +12,14 @@ jobs:
       - uses: dtolnay/rust-toolchain@master
         with:
           components: rustfmt
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-09-21
       - uses: Swatinem/rust-cache@v2
       - name: Install cargo-llvm-cov
         uses: taiki-e/install-action@cargo-llvm-cov
         # TODO: Merge coverage reports for tests on different architectures.
         # <https://github.com/taiki-e/cargo-llvm-cov?tab=readme-ov-file#merge-coverages-generated-under-different-test-conditions>
       - name: Generate code coverage
-        run: cargo +nightly-2024-01-04 llvm-cov --codecov --output-path codecov.json
+        run: cargo +nightly-2024-09-21 llvm-cov --codecov --output-path codecov.json
         env:
           RUSTFLAGS: "-C target-feature=+avx512f"
       - name: Upload coverage to Codecov

crates/prover/benches/fft.rs

@@ -29,7 +29,7 @@ pub fn simd_ifft(c: &mut Criterion) {
                 || values.clone().data,
                 |mut data| unsafe {
                     ifft(
-                        transmute(data.as_mut_ptr()),
+                        transmute::<*mut PackedBaseField, *mut u32>(data.as_mut_ptr()),
                         black_box(&twiddle_dbls_refs),
                         black_box(log_size as usize),
                     );
@@ -58,7 +58,7 @@ pub fn simd_ifft_parts(c: &mut Criterion) {
             || values.clone().data,
             |mut values| unsafe {
                 ifft_vecwise_loop(
-                    transmute(values.as_mut_ptr()),
+                    transmute::<*mut PackedBaseField, *mut u32>(values.as_mut_ptr()),
                     black_box(&twiddle_dbls_refs),
                     black_box(9),
                     black_box(0),
@@ -72,7 +72,7 @@ pub fn simd_ifft_parts(c: &mut Criterion) {
             || values.clone().data,
             |mut values| unsafe {
                 ifft3_loop(
-                    transmute(values.as_mut_ptr()),
+                    transmute::<*mut PackedBaseField, *mut u32>(values.as_mut_ptr()),
                     black_box(&twiddle_dbls_refs[3..]),
                     black_box(7),
                     black_box(4),
@@ -91,7 +91,7 @@ pub fn simd_ifft_parts(c: &mut Criterion) {
             || transpose_values.clone().data,
             |mut values| unsafe {
                 transpose_vecs(
-                    transmute(values.as_mut_ptr()),
+                    transmute::<*mut PackedBaseField, *mut u32>(values.as_mut_ptr()),
                     black_box(TRANSPOSE_LOG_SIZE as usize - 4),
                 )
             },
@@ -115,8 +115,10 @@ pub fn simd_rfft(c: &mut Criterion) {
             target.set_len(values.data.len());
 
             fft(
-                black_box(transmute(values.data.as_ptr())),
-                transmute(target.as_mut_ptr()),
+                black_box(transmute::<*const PackedBaseField, *const u32>(
+                    values.data.as_ptr(),
+                )),
+                transmute::<*mut PackedBaseField, *mut u32>(target.as_mut_ptr()),
                 black_box(&twiddle_dbls_refs),
                 black_box(LOG_SIZE as usize),
             )

crates/prover/benches/merkle.rs

@@ -21,7 +21,7 @@ fn bench_blake2s_merkle<B: MerkleOps<Blake2sMerkleHasher>>(c: &mut Criterion, id
     let n_elements = 1 << (LOG_N_COLS + LOG_N_ROWS);
     group.throughput(Throughput::Elements(n_elements));
     group.throughput(Throughput::Bytes(N_BYTES_FELT as u64 * n_elements));
-    group.bench_function(&format!("{id} merkle"), |b| {
+    group.bench_function(format!("{id} merkle"), |b| {
         b.iter_with_large_drop(|| B::commit_on_layer(LOG_N_ROWS, None, &col_refs))
     });
 }

crates/prover/src/constraint_framework/component.rs

@@ -55,9 +55,10 @@ impl TraceLocationAllocator {
 }
 
 /// A component defined solely in means of the constraints framework.
+///
 /// Implementing this trait introduces implementations for [`Component`] and [`ComponentProver`] for
-/// the SIMD backend.
-/// Note that the constraint framework only support components with columns of the same size.
+/// the SIMD backend. Note that the constraint framework only supports components with columns of
+/// the same size.
 pub trait FrameworkEval {
     fn log_size(&self) -> u32;
 

crates/prover/src/core/air/accumulation.rs

@@ -1,4 +1,5 @@
 //! Accumulators for a random linear combination of circle polynomials.
+//!
 //! Given N polynomials, u_0(P), ... u_{N-1}(P), and a random alpha, the combined polynomial is
 //! defined as
 //!   f(p) = sum_i alpha^{N-1-i} u_i(P).

crates/prover/src/core/air/mod.rs

@@ -15,10 +15,10 @@ mod components;
 pub mod mask;
 
 /// Arithmetic Intermediate Representation (AIR).
-/// An Air instance is assumed to already contain all the information needed to
-/// evaluate the constraints.
-/// For instance, all interaction elements are assumed to be present in it.
-/// Therefore, an AIR is generated only after the initial trace commitment phase.
+///
+/// An Air instance is assumed to already contain all the information needed to evaluate the
+/// constraints. For instance, all interaction elements are assumed to be present in it. Therefore,
+/// an AIR is generated only after the initial trace commitment phase.
 // TODO(spapini): consider renaming this struct.
 pub trait Air {
     fn components(&self) -> Vec<&dyn Component>;

crates/prover/src/core/backend/cpu/quotients.rs

@@ -75,10 +75,10 @@ pub fn accumulate_row_quotients(
     row_accumulator
 }
 
-/// Precompute the complex conjugate line coefficients for each column in each sample batch.
-/// Specifically, for the i-th (in a sample batch) column's numerator term
-/// `alpha^i * (c * F(p) - (a * p.y + b))`, we precompute and return the constants:
-/// (`alpha^i * a`, `alpha^i * b`, `alpha^i * c`).
+/// Precomputes the complex conjugate line coefficients for each column in each sample batch.
+///
+/// For the `i`-th (in a sample batch) column's numerator term `alpha^i * (c * F(p) - (a * p.y +
+/// b))`, we precompute and return the constants: (`alpha^i * a`, `alpha^i * b`, `alpha^i * c`).
 pub fn column_line_coeffs(
     sample_batches: &[ColumnSampleBatch],
     random_coeff: SecureField,
@@ -103,8 +103,9 @@ pub fn column_line_coeffs(
         .collect()
 }
 
-/// Precompute the random coefficients used to linearly combine the batched quotients.
-/// Specifically, for each sample batch we compute random_coeff^(number of columns in the batch),
+/// Precomputes the random coefficients used to linearly combine the batched quotients.
+///
+/// For each sample batch we compute random_coeff^(number of columns in the batch),
 /// which is used to linearly combine the batch with the next one.
 pub fn batch_random_coeffs(
     sample_batches: &[ColumnSampleBatch],

crates/prover/src/core/backend/simd/bit_reverse.rs

@@ -159,7 +159,7 @@ mod tests {
 
         let res = bit_reverse16(values.data.try_into().unwrap());
 
-        assert_eq!(res.map(PackedM31::to_array).flatten(), expected);
+        assert_eq!(res.map(PackedM31::to_array).as_flattened(), expected);
     }
 
     #[test]

crates/prover/src/core/backend/simd/blake2s.rs

@@ -369,8 +369,12 @@ mod tests {
 
         let res_vectorized: [[u32; 8]; 16] = unsafe {
             transmute(untranspose_states(compress16(
-                transpose_states(transmute(states)),
-                transpose_msgs(transmute(msgs)),
+                transpose_states(transmute::<Aligned<A64, [[u32; 8]; 16]>, [u32x16; 8]>(
+                    states,
+                )),
+                transpose_msgs(transmute::<Aligned<A64, [[u32; 16]; 16]>, [u32x16; 16]>(
+                    msgs,
+                )),
                 u32x16::splat(count_low),
                 u32x16::splat(count_high),
                 u32x16::splat(lastblock),

crates/prover/src/core/backend/simd/circle.rs

@@ -85,10 +85,7 @@ impl SimdBackend {
 
     // Generates twiddle steps for efficiently computing the twiddles.
     // steps[i] = t_i/(t_0*t_1*...*t_i-1).
-    fn twiddle_steps<F: Field>(mappings: &[F]) -> Vec<F>
-    where
-        F: FieldExpOps,
-    {
+    fn twiddle_steps<F: Field + FieldExpOps>(mappings: &[F]) -> Vec<F> {
         let mut denominators: Vec<F> = vec![mappings[0]];
 
         for i in 1..mappings.len() {
@@ -151,7 +148,7 @@ impl PolyOps for SimdBackend {
         // Safe because [PackedBaseField] is aligned on 64 bytes.
         unsafe {
             ifft::ifft(
-                transmute(values.data.as_mut_ptr()),
+                transmute::<*mut PackedBaseField, *mut u32>(values.data.as_mut_ptr()),
                 &twiddles,
                 log_size as usize,
             );
@@ -254,8 +251,8 @@ impl PolyOps for SimdBackend {
             // FFT from the coefficients buffer to the values chunk.
             unsafe {
                 rfft::fft(
-                    transmute(poly.coeffs.data.as_ptr()),
-                    transmute(
+                    transmute::<*const PackedBaseField, *const u32>(poly.coeffs.data.as_ptr()),
+                    transmute::<*mut PackedBaseField, *mut u32>(
                         values[i << (fft_log_size - LOG_N_LANES)
                             ..(i + 1) << (fft_log_size - LOG_N_LANES)]
                             .as_mut_ptr(),

crates/prover/src/core/backend/simd/fft/ifft.rs

@@ -589,7 +589,7 @@ mod tests {
         let mut res = values;
         unsafe {
             ifft3(
-                transmute(res.as_mut_ptr()),
+                transmute::<*mut PackedBaseField, *mut u32>(res.as_mut_ptr()),
                 0,
                 LOG_N_LANES as usize,
                 twiddles0_dbl,
@@ -655,7 +655,7 @@ mod tests {
             [val0.to_array(), val1.to_array()].concat()
         };
 
-        assert_eq!(res, ground_truth_ifft(domain, values.flatten()));
+        assert_eq!(res, ground_truth_ifft(domain, values.as_flattened()));
     }
 
     #[test]
@@ -669,7 +669,7 @@ mod tests {
             let mut res = values.iter().copied().collect::<BaseColumn>();
             unsafe {
                 ifft_lower_with_vecwise(
-                    transmute(res.data.as_mut_ptr()),
+                    transmute::<*mut PackedBaseField, *mut u32>(res.data.as_mut_ptr()),
                     &twiddle_dbls.iter().map(|x| x.as_slice()).collect_vec(),
                     log_size as usize,
                     log_size as usize,
@@ -691,11 +691,14 @@ mod tests {
             let mut res = values.iter().copied().collect::<BaseColumn>();
             unsafe {
                 ifft(
-                    transmute(res.data.as_mut_ptr()),
+                    transmute::<*mut PackedBaseField, *mut u32>(res.data.as_mut_ptr()),
                     &twiddle_dbls.iter().map(|x| x.as_slice()).collect_vec(),
                     log_size as usize,
                 );
-                transpose_vecs(transmute(res.data.as_mut_ptr()), log_size as usize - 4);
+                transpose_vecs(
+                    transmute::<*mut PackedBaseField, *mut u32>(res.data.as_mut_ptr()),
+                    log_size as usize - 4,
+                );
             }
 
             assert_eq!(res.to_cpu(), ground_truth_ifft(domain, &values));