[EPIC] Benchmarks #2288

Goal

Rollkit v1 had a lot of simplification and a introduction of the execution API. After the testing and reliability epic is completed we should provide benchmarks for the Cosmos SDK and EVM execution in order to display to users how well the system works.

Description

We should benchmark and identify areas of slow down in the code so that we can fix and improve the system. On top of the improvements of rollkit we should strive to optimise the performance of the execution environments.

Benchmarks should cover three areas:

• Ev-Node
• ABCI Execution (Cosmos SDK)
• EVM execution (Reth)

Related Tasks

• [FEATURE] Add granular tracing instrumentation with #[instrument] ev-reth#137 Add granular tracing instrumentation with #[instrument]

  • feat: adding additional fine grained spans ev-reth#142

• Benchmark rollkit + testapp against node and app. (this is used to uncover rollkit performance bottlenecks)

  • [FEATURE] Add testapp throughput benchmark to isolate ev-node overhead #3130

• Benchmark rollkit + ABCI Execution (Cosmos SDK) in order to provide potential performance numbers

• Benchmark rollkit + EVM Execution (RETH) in order to provide potential performance numbers

  • feat(benchmarking): adding ERC20 benchmarking test #3114
  • feat(benchmarking): adding gas burner test #3115

• fix(block): fix blocktime timer usage #3137

Dependencies and Risks

• Dependencies:

  • Celestia Node
  • Celestia App
  • RETH
  • Cosmos SDK

• Risks:

  • Performance is not what is expected

Notes and References

No response