# ignacio (jsign) - Update 4

This week I worked on refining and extending a set of benchmarks about hot-path functions comparing Merkle Patricia Tries with Verkle Tries:

- Refining Keccak benchmarks.
- Comparing the Go cryptography library implementations to be used by Geth against:
- Go BLST (CGO)
- Native BLST (C)
- Arkworks (Rust)
- Compare hashing keys to construct the trie:
- Keccak (used in Merkle Patricia Tries)
- Pedersen Hash (used in Verkle Tries)
- Make all these benchmarks easily verifiable with a single command.
- Report the results under two different setups.
- [Discovered why](https://discord.com/channels/595666850260713488/824798757618188339/1037123672265216000) the Go `go-ipa` library is incredibly fast for 255 bit finite-field multiplication (using Intel ADX).
- [Found a small bug](https://discord.com/channels/595666850260713488/824798757618188339/1036016586089299978) in the current Geth implementation compared to the EIP while creating trie keys for Verkle Tries.

All this work is happening in an [open-source repo](https://github.com/jsign/verkle-vs-patricia), you can refer to the PRs to see more details:

- https://github.com/jsign/verkle-vs-patricia/pull/1
- https://github.com/jsign/verkle-vs-patricia/pull/2
- See repo readme.

You can also find a lot of discussion around this work with Dankrad, Kevaundray and Guillaume in the #verkle-trie-migration Discord channel.