Originally posted by **btilmon** October 27, 2023
I saw Niall Emmart's talk from the gpu masterclass on his Goldilocks NTT approach (code [here](https://github.com/yrrid/GoldilocksNTT)). Would his approach be faster than the current icicle implementation of NTT? Considering opening a PR here to add support for it.
Discussed in https://github.com/ingonyama-zk/icicle/discussions/252