Background

zkEVM development is accelerating, plus the plan to raise the gas limit, and the updated EIP process now mandates benchmarks for every new opcode/precompile. Multiple teams might have begun writing their own benchmark tests, which might lead to duplicated work and incompatible infrastructure.

Related issue: #1704

Proposal

Move all benchmark tests, and the tooling that runs them, into EEST. This is already a consensus during Berlin interop.
We need to further add a dedicated “Benchmarks” section to the EEST docs that provides the following features and functionality.

This is my personal idea, any feedback and suggestion is welcome:

Opcode coverage table

• Lists every opcode alongside the benchmark tests that exercise it.
• Highlights missing coverage (e.g. PC is still un-benchmarked; We missed it even though we'd created Create blockchains that consume full gas on single operations. #1453 ).
• Example: selfdestruct - test_worst_selfdestruct_existing, test_worst_selfdestruct_initcode, test_worst_selfdestruct_created

Concise test descriptions

• Short and clear description, so contributors grasp each test’s intent without reading code

Results & resources hub

• Links to the latest benchmark dashboards, raw results.

Collaboration channel

• A lightweight forum (issue template or discussion board) where teams can propose new benchmarks, coordinate on shared scenarios, and avoid duplicating effort. As the testing scenario will be more and more complicated, such as the one in Issue tracker for porting extra benchmark tests #1923

Implementing this structure will give contributors a clear view of existing coverage, streamline future additions, and keep benchmark infrastructure consistent across the ecosystem.