Releases: Consensys/gnark-crypto
Releases · Consensys/gnark-crypto
v0.17.0
What's Changed
- fix: missing Poseidon2 round keys by @Tabaie in #621
- feat: Poseidon2 Hash Instantiation for BLS12-377 by @Tabaie in #623
- feat: add Grumpkin elliptic curve (2-cycle with BN254) by @yelhousni in #625
- Perf: Poseidon2 GKR circuit by @Tabaie in #628
- feat: add sis avx512 and fft avx512 for koalabear by @gbotrel in #622
- InterpolateOnRange refactor by @Tabaie in #634
- chore: add auto close PR workflow by @gbotrel in #638
- perf: subgroup membership by @yelhousni in #635
- feat: poseidon2 for koala-bear, baby-bear and goldilocks by @yelhousni in #629
- chore: generify poseidon2 parameters for other curves/fr by @yelhousni in #636
- feat: baby-bear and koala-bear extensions of degree 4 by @yelhousni in #643
- fix: ensure fast path is taken only with fixed bound and degree by @gbotrel in #651
- Poseidon2 compression for small fields by @Tabaie in #644
- fix: poseidon2 templates by @yelhousni in #648
- test: improve NAF decomposition test coverage by @DeVikingMark in #617
- refactor: generate code for poseidon2_test by @yelhousni in #660
- test(bn254, bls12-381): test points intentionally not on sugroups G1/2 by @yelhousni in #658
- refactor: generify small fields extensions by @yelhousni in #647
New Contributors
- @DeVikingMark made their first contribution in #617
Full Changelog: v0.16.0...v0.17.0
Contributors
gbotrel, Tabaie, and 2 other contributors
Assets 2
v0.16.0
v0.15.0
What's Changed
Features
- Eisenstein integers (#543).
- Poseidon2 (#553).
- BabyBear and KoalaBear fields: 31-bit vanilla Go implementation (#558, #577).
- FFLonK (#498).
- KZG MPC setup (#589).
- Hash registry with state storer (#564).
- Miller loop with direct sextic extension (BW6-761) (#608).
Performance Improvements
- Add assembly implementations: Vector::Add, Sub, and ScalarMul (purego included) (#536).
- AVX512 optimizations: vector.Sum, vector.InnerProduct + assembly refactor (#547).
- AVX512 vector operations: KoalaBear and BabyBear fields (#568).
- ARM64 assembly optimizations: 4 and 6 words modulus (#554).
- ARM NEON assembly: KoalaBear and BabyBear fields (#588).
- Subgroup check optimizations (#557).
- SetBytes performance: Avoids allocations in non-fast paths (#599).
Fixes
- Polynomial evaluation (#539).
- Shplonk Buffer sizing (#551).
- Typos (#562, #610).
- Comment and variable naming fixes (#603).
Refactors
- Removed curves: BLS12-378 and BW6-756 (#534).
- SIS cleanup (#596).
- Removed ECC Encoder dependency: FFT domain marshal (#565).
- Cleaned IOP (#587).
Documentation
- Updated README (#533).
- Documented batchAddAffine function (#555).
- Updated element package documentation (#572).
- Less verbose Apache 2 header (#570).
- Fixed non-generated headers (#571).
- Changed copyright year to 2025 (#595).
Full Changelog: v0.14.0...v0.15.0
v0.14.0
What's Changed
- refactor: correct comments by @yelhousni in #511
- Feat/pedersen batch verify by @Tabaie in #517
- build: update runner and go version by @gbotrel in #531
- perf: remove uneeded bucket init in affine coordinates fixes #529 by @gbotrel in #530
- feat: Pedersen vector commitment scheme multi VK verification by @ivokub in #527
- chore: fix some function names by @drawdrop in #526
- fix: fixes #522 with bound check before computing twiddles when domain has no precompute set by @gbotrel in #523
- refactor(bls12-381, bls12-377): MillerLoopFixed corresponds to gnark by @yelhousni in #524
New Contributors
Full Changelog: v0.13.0...v0.14.0
Contributors
ivokub, gbotrel, and 3 other contributors
Assets 2
v0.13.0
What's Changed
- perf, feat, refactor: improve some plonk primitives by @gbotrel in #456
- ci: update ci workflows by @gbotrel in #460
- Feat/kzg dummy srs by @ThomasPiellard in #458
- Perf: faster final exponentiation (hard part) for BW6 curves by @yelhousni in #459
- feat: wrap existing hash-to-field for implementing hash.Hash interface by @ivokub in #464
- Refactor(BW6): use revisited Ate pairing instead of Tate by @yelhousni in #465
- fix: remove duplicate gkr template generation by @ivokub in #439
- fix: edge case where domain cardinality is 1 by @gbotrel in #454
- Perf: KZG with fixed-argument pairing in affine coordinates by @yelhousni in #466
- Refactor: re-introduce points in KZG verification key by @yelhousni in #469
- Add SIS to BLS12-377 by @AlexandreBelling in #470
- chore(deps): bump golang.org/x/crypto from 0.10.0 to 0.17.0 by @dependabot in #473
- feat: allow hashing Fiat-Shamir transcript challenge without hashing to field first by @ivokub in #474
- perf,memory: fft without precompute by @gbotrel in #472
- refactor: kzg srs template by @yelhousni in #475
- doc: update hash to curve rfc ref and impl comment by @hussein-aitlahcen in #478
- feat: adds byteorder option in mimc
Newby @gbotrel in #480 - perf: optimize curve membership test by @yelhousni in #487
- refactor: make pedersen vk fields public by @ahmetyalp in #488
- fix: gkr add gate evaluate by @BeratOz01 in #490
- fix: fixes #491 by @ThomasPiellard in #492
- Chores: export the unrolled FFT64 out of SIS by @AlexandreBelling in #496
- feat: ECDSA error types by @ivokub in #497
- perf: add kzg.UnsafeToBytes and kzg.UnsafeFromBytes methods by @gbotrel in #499
- refactor: move test utils package into isolation to avoid every user to import test stuff by @gbotrel in #500
- perf: unsafe.WriteSlice (#501) by @gbotrel in #503
- perf(ecc): optimize affine Add, Sub and Double by @yelhousni in #509
- perf(ecc): faster affine Add by @yelhousni in #510
- perf: KZG verification uses
JointScalarMultiplicationby @yelhousni in #493 - fix: sis limb-decomposition works with log-two_bound > 8 by @AlexandreBelling in #512
- refactor: remove E3 Conjugate from BW6 by @yelhousni in #514
- perf(bn254): optimize affine pairing with doubleAndAdd by @yelhousni in #506
- fix: flaky FRI test by @ivokub in #518
- docs: describe length extension attack for MiMC hash function. by @ivokub in #520
- fix(bandersnatch): GLV bounds + test by @yelhousni in #516
New Contributors
- @dependabot made their first contribution in #473
- @ahmetyalp made their first contribution in #488
- @BeratOz01 made their first contribution in #490
Full Changelog: v0.12.1...v0.13.0
Contributors
ivokub, hussein-aitlahcen, and 7 other contributors
Assets 2
v0.12.1
What's Changed
- fix: GLV exponentiation in GT with large exponents by @ThomasPiellard in #451 Security Advisory
- feat: kzg extended transcript by @ThomasPiellard in #452
- perf(fft): introduce cache efficient bit reverse shuffling by @gbotrel in #446
- perf: Improve performance of point equality checks by @jsign in #450
Full Changelog: v0.12.0...v0.12.1
Contributors
jsign, gbotrel, and ThomasPiellard
Assets 2
v0.12.0
What's Changed
-
fix malleability sig by @ThomasPiellard in #449 https://nvd.nist.gov/vuln/detail/CVE-2023-44273
-
perf: multiexp, avoid direct coordinate access to check for zero points by @jsign in #414
-
perf: edwards, improve the performance of Add, MixedAdd and IsOnCurve by @jsign in #441
-
perf: edwards, avoid inversions in Add in extended points by @jsign in #442
Full Changelog: v0.11.2...v0.12.0
Contributors
jsign, gbotrel, and ThomasPiellard
Assets 2
v0.11.2
What's Changed
- Fix some typos by @jtraglia in #394
- Adding testing for deserialization of G1 and G2 points by @asanso in #393
- Fix some implicit memory aliasing in for loops by @jtraglia in #395
- Do not XOR with zero by @jtraglia in #398
- Disable check shadowing in govet linter by @jtraglia in #397
- Add a bunch of "nosec G404" comments in test code by @jtraglia in #399
- Enable misspell linter & fix findings by @jtraglia in #401
- Fix
RSis.CopyWithFreshBufferby @AlexandreBelling in #402 - feat: Marshal [][]fr.Element by @Tabaie in #400
- Run golangci-lint on generated files by @jtraglia in #396
- docs: ConsenSys -> Consensys by @Tabaie in #406
- msm: semaphore to limit CPUs + better split strategy (up to 25% perf boost on 96cores) by @gbotrel in #403
- Feat/fold pedersen by @Tabaie in #407
- fix: do not read empty slices as nil by @Tabaie in #410
- fix: incorrect semaphore init could cause msm deadlock by @gbotrel in #411
- edwards: optimize point negation by @jsign in #413
- Feat/gkr custom gates by @Tabaie in #419
- perf: fast path for SIS with logTwoBound: 8, logTwoDegree: 6 by @gbotrel in #416
- feat: add WriteRawTo, UnsafeReadFrom to kzg.ProvingKey by @gbotrel in #422
- Fix/gkr eq bug by @Tabaie in #421
- feat: add AsyncReadFrom to fr.Vector and fft.Domain by @gbotrel in #424
- fix: ECDSA HashToInt bytes-bits mismatch by @ivokub in #428
- Small optimization over the memory usage of MiMC by @AlexandreBelling in #435
- perf: improve fft domain memory footprint by @gbotrel in #437
- Refactor/gkr test vectors by @Tabaie in #425
- v0.11.2 by @gbotrel in #438
New Contributors
Full Changelog: v0.11.1...v0.11.2
Contributors
asanso, ivokub, and 5 other contributors
