Skip to content

Upstream EVM benchmarking #22

@chfast

Description

@chfast

The current scripts for running benchmarks on EVM implementations are in https://github.com/cdetrio/benchmarking-wasm-ewasm-evm (private repo).

  • The scripts could be moved to this repo.

They are based on out-of-tree changes that should be upstreamed to:

Also consider:

  • Building individual docker images for each EVM instead of having single one.

Recent results: https://docs.google.com/presentation/d/1MgXd159bmx37Q70Ng_ggHcClcYLafYqrxNEBFKAMqWI/edit#slide=id.g64bd7b8151_0_1

  • evmone results from docker build and from native build are different - the difference is big enough that only usage of different compiler version is not enough to explain it.

The EVM should accept for benchmarking

  • EVM bytecode
  • EVM input

From each EVM we need following data:

  • Total time of execution
  • EVM output (to check correctness)
  • gas used
  • EVM exception/success status (not strictly needed)

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type
No fields configured for issues without a type.

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions