Refactor(BW6-761): use revisited Ate pairing instead of Tate#876
Conversation
|
📦 ❌ 📦 ❌ 📦 ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ |
yelhousni
changed the title
|
📦 ❌ 📦 ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ |
|
@ivokub this is ready for review after we merge Consensys/gnark-crypto#465 |
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
ivokub
left a comment
ivokub
left a comment
There was a problem hiding this comment.
Actually, I was relooking and there were some errors. Will submit another review after testing and recompiling
There was an old test of a function that does not exist anymore ( |
|
📦 ❌ 📦 ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ 📦 ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ ❌ |
ivokub
left a comment
ivokub
left a comment
There was a problem hiding this comment.
Yup, running the KZG and Groth16 examples also timeouted my other PR. And the test with Mul034 was the failing one. Now looks perfect!
2 participants
Description
Revisit the Miller loop algorithm from #846 to allow fixed-argument pairings needed for KZG optimization (see #874). TL;DR: use Eq. (6') instead of (6) in https://hackmd.io/@gnark/BW6-761-changes.
The PR needs gnark-crypto PR Consensys/gnark-crypto#465 for testing.
TODO:
Type of change
How has this been tested?
Same existing tests work fine + new tests for fixed-argument pairing (compare pairing result to gnark-crypto).
How has this been benchmarked?
We lost 2084 R1CS and 24527 SRS for a single pairing compared to
masterbut this should be offset in KZG gadget #874 with the fixed-argument pairing that the new Miller loop allows (see below).Checklist:
golangci-lintdoes not output errors locally