Reth is now faster, smaller, and ready for the future of crypto infrastructure.
Gigagas per second has not just been achieved, it's been blown out of the water.
We've put in a ton of work in this, and we're incredibly proud to be sharing this with everyone.
Reth 2.0 is really incredible and not just on small Ethereum-size blocks.
We took benchmarking seriously, built synthetic benchmarks for gigagas blocks, and we measured everything in depth.
We had to improve our infrastructure to reduce the time to run an E2E benchmark from hours to minutes. This required insane amounts of work & tuning.
We even built an auto-optimization harness similar to Karpathy's autoresearch which we hope to get a lot of gains in the coming weeks.
We got on this performance journey around 2024, when we pushed the industry towards aligning around "Gigagas per second" as the goal.
It took us over a year to tame the "parallel, pipelined, modified state root", but we did it, without having to resort to a trie-specific database AND preserving the amazing archive node capabilities of Reth.
We haven't yet the fundamental limits of what we can do.
Ethereum should reach the efficient frontier of what’s possible before arguing about how we would choose between our values once we’ve hit those limits.
Shipping faster will help Ethereum get there.
Instead of "should we do X or Y?", the answer might be "both."
We can move faster by deciding to evolve the protocol with ambition, and not ossify.
We can do better than N-of-N (5 CL + 5 EL!) clients coming to consensus before new upgrades.
We can level up ACD.
We can allocate more resources to Devops & Testing.
We started @Ithacaxyz to accelerate the frontier, & have raised $20M from @paradigm.
We’ve been collaborating with the developer community on some of crypto's hardest problems. Small teams = big impact. The future of crypto will be built together.
Over the past four years, our team shipped some of the industry’s most used open source tools like Reth & Foundry.
We believe in sustainable open source development for the good of crypto.
We think the stack is ready, and it’s now time to accelerate.
After almost two years of development and a successful audit by Sigma Prime, we are finally releasing Reth 1.0, the first “prod-ready” release of our blazing-fast Ethereum execution client. We invite RPC providers and stakers to run Reth.
To speed up execution by getting rid of the overheads of bytecode interpretation, inspired by well-established technologies like the Java or the WASM JIT.