Feature Request: Ethereum Classic Support with SHA3 Proof of Work

[antsankov]

Hello,

My name is Alex and I am working on ECIP-1049! This ECIP changes the Proof of Work of the Ethereum Classic network from Ethash to SHA3 (Keccak256). SHA3 is capable of much higher block validation and hashing speeds than Ethash, and can greatly reduce the complexity of nodes and mining hardware. It is a Proof of Work scaling solution.

Ethash has also not worked in being ASIC-resistant, and the community has decided to move on from it since it is now technical debt that requires excess memory to mine (3.2GB) and a pseudo-random cache to validate (50+ MB just to calculate)

Peer-reviewed analysis has shown SHA3 has over 4x the efficiency per hash over Bitcoin SHA2:

So far we have created a SHA3 testnet, called "Astor Network" which was made from a forked Parity Ethereum client. This network is ETH based, and can be found here. We also have created an proof of concept CPU Toy Miner.

Is Turbo-geth interested in supporting this network?

Extra Materials:

• Astor Testnet Site
• Talk I gave to the ETC Community in Vancouver in 2019

[AlexeyAkhunov]

Hi Alex, thank you for stopping by! I have been following some of your work on SHA3 mining online, and heard of the test net.
We did think about the ETC support, and I am sure it would be possible, but it is not currently our priority. Our priority is to make turbo-geth fully functional and "ship" it, then produce documentation on how other implementations can migrate to our data model, which I believe offers superior trade-offs.

We would definitely be interested in contributions that would enable turbo-geth to connect to ETC network, but we cannot currently accommodate this work ourselves.
Hope this helps, Alexey

[antsankov]

Thank you for the support Alexey! Good luck with the work, I will close the issue and monitor the turbo-geth project. Once an official release is made, I will be happy to revisit and possibly implement changes myself.

Before I close, I had one question: Is Turbo-geth planning on being an Eth1.x and Eth2.0 client or just Eth1.x exclusively?

[AlexeyAkhunov]

Thank you! Only Eth1.x exclusively

[bobsummerwill]

Hey guys!

So the ETC Cooperative will be funding Keccak256 implementation in Hyperledger Besu, Parity-Ethereum and MultiGeth.

Peter won't take anything upstream into Geth from ETC, but we should have a PR against MultiGeth which should "just work" for TurboGeth as well.