From c7e2dbdad912749c128caa588a76b5dd59bb0d64 Mon Sep 17 00:00:00 2001 From: Justin Thaler <39494992+GUJustin@users.noreply.github.com> Date: Sat, 5 Oct 2024 08:05:17 -0400 Subject: [PATCH] Update folding.md --- book/src/future/folding.md | 2 -- 1 file changed, 2 deletions(-) diff --git a/book/src/future/folding.md b/book/src/future/folding.md index 0e955b43c..797042a31 100644 --- a/book/src/future/folding.md +++ b/book/src/future/folding.md @@ -55,6 +55,4 @@ This time overhead can be reduced with additional engineering/optimizations. For
  • For at least the first round of Spartan's sum-check, we could store only $32$ bits per entry of $a, b, c$ rather than 256, which could save another factor of two in space cost. Also, prover speed optimizations for applying sum-check to small values described in Bagad-Domb-Thaler can also lead to additional space improvements for Spartan. At some point, the space cost of computing grand product proofs would dominate the space cost of Spartan proving and it won't be worthwhile to cut Spartan proving space further.
  • -
  • Using HyperNova in place of Nova will avoid committing to cross terms consisting of random field elements, which would cut the time cost of recursive proving by an order of magnitude. But HyperNova implementations are not yet as mature as Nova itself.
  • -