v0.0.5

What's Changed

• fix power gate by @zhenfeizhang in #128
• Recursive Verifier by @zhiyong1997 in #127
• Lagrange Form Bi kzg update by @zhiyong1997 in #130
• Adding sumcheck-cuda by @slzk in #131
• stats code & release binary by @zhiyong1997 in #135
• fix-binary-release by @zhiyong1997 in #136
• Reuse eq_evals in prover, verifier, and recursive verifier by @zhiyong1997 in #138
• pcs trait v1 by @zhiyong1997 in #140

Full Changelog: v0.0.4...v0.0.5

v0.0.4

Expander v0.0.4 Release note

Summary

• Major changes:
  • Major code refactoring with modularization
  • Add support for MPI
  • Completed GF2_ext128 field
  • Implemented GKR^2 prover
  • Allow for public inputs at arbitrary layers
  • Optimization of proof size: substitute the SIMD field with the native field in the proof message
• Minor changes:
  • Improved test coverage
  • Automated end to end tests
  • Improved CI

What's Changed

• Zz/refactor and clean up by @zhenfeizhang in #55
• Ci benchmark by @niconiconi in #57
• Automate benchmark by @niconiconi in #58
• Minor: Try to resolve some TODOs in neon by @tonyfloatersu in #60
• Feat/criterion by @niconiconi in #61
• Update circuit.rs by @zhiyong1997 in #63
• fix performance loss by @zhenfeizhang in #66
• allow for different FiatShamir hashes by @zhenfeizhang in #65
• Minor: CI process simplify with dev-setup loading circuits by @tonyfloatersu in #62
• Minor: Serialization and Deserialization Error Propagation by @tonyfloatersu in #64
• Add sumcheck single iteration for CUDA sumcheck development by @slzk in #69
• Zz/gf2 128 mul by x by @zhenfeizhang in #68
• Revert "Zz/gf2 128 mul by x" by @niconiconi in #70
• Squash merge GF2_dev into main, accepting all GF2_dev changes by @niconiconi in #79
• Sumcheck opt by @niconiconi in #81
• handle layer with zero var. by @zhiyong1997 in #84
• Implement more field operations for ECC by @mcfx in #86
• avx256 by @chonpsk in #88
• Proof size by @zhiyong1997 in #87
• fix_pack by @niconiconi in #89
• Verification time by @zhiyong1997 in #90
• update CI by @zhenfeizhang in #102
• remove avx512 instructions from avx256 by @mcfx in #94
• [FIX] m31 exp by @enpsi20481 in #96
• fix ci by @zhenfeizhang in #104
• Mpi by @zhiyong1997 in #93
• use avx256 by default by @zhenfeizhang in #109
• Update readme.md by @siq1 in #110
• refactoring field impl by @zhenfeizhang in #95
• transcript refactor by @zhenfeizhang in #103
• Public input rebase by @zhiyong1997 in #108
• refactor sumcheck by @zhenfeizhang in #112
• code refactor by @zhenfeizhang in #115
• Zz/fix poseidon witnesses by @zhenfeizhang in #118
• Zz/fix script by @zhenfeizhang in #120
• Mpi chunk size by @zhiyong1997 in #121
• Mpi chunk size by @zhiyong1997 in #123
• Fix expander-exec by @mcfx in #122
• fix mpi by @zhiyong1997 in #124
• Fix scratchpad and Optimize Expander Eq by @zhiyong1997 in #125

New Contributors

• @slzk made their first contribution in #69
• @mcfx made their first contribution in #86
• @chonpsk made their first contribution in #88
• @enpsi20481 made their first contribution in #96
• @siq1 made their first contribution in #110

Full Changelog: v0.0.3...v0.0.4

v0.0.3

Add a ready checking entrypoint #54

v0.0.2

We fixed the soundness issue by Justin Thaler and we did a lot of optimizations to match our old performance number. Now it matches with the old buggy version.

v0.0.1

enable debug logging (#28)