Real trading data makes the chain faster, and RISE VM is now 2x faster than revm on production blocks.
A while back, we decided to downstream revm into our own RISE VM implementation for better API control, specifically for better cache management and Parallel EVM "build-up".
Along the way, we also realized how non-optimal upstream was, and after some register optimizations, op-code fusions, smart inlining and caching (thanks to gigatons of @risextrade data to analyze),
Our early RISE VM adaption has been consistently processing @risechain blocks 2x faster than revm!
As we push for even more aggressive optimizations, and build up Parallel EVM on top of this tight core, I don't think any EVM chain can get within 10% of our performance.
Towards 1000 on-chain orderbooks!?View quoted post on X
