At the end of the BCH dev meeting (2020 meeting #1), I presented a rough draft of a series of benchmark tests that I'm developing to run against BCH infrastructure. The tests are based on the concept of the Cash Stack, which is a stack of software that includes a full node, indexer, REST API, SDK, and application (like a wallet).
The idea is to test the system-as-a-whole, but variations of the test stress specific sub-components, like the indexer or full node. The goal is to create standardized tests that can be applied to different configurations, and allow an apples-to-apples comparison between different full node implementation, indexers, etc.
I've released v1.0.0 of the benchmark tests with some initial baseline data. I'd love for developers to review the documents in the repository and provide any critical feedback. Feedback can be left as a GitHub Issue in the repository:
https://github.com/christroutner/benchmark-bch/tree/master/docs
I invite organizations behind the various full node implementations and indexers to work with me. It would be great to benchmark the different pieces of infrastructure against one another. I've already reached out to OpenBazaar to test the BCH indexer behind their infrastructure. I'll also be running these benchmark tests against SLPDB and the Bitcoin.com SLP Java Indexer.
I would more like to see a test to measure transaction signing capacity. Like, generating 100k new address, creating and signing a transaction for each of the address (but not broadcasting anything to avoid actual spending), then deleting all of these address from the wallet. Measuring how long this will take, measuring how much ram will the wallet consume, will it pass or actually crash... This would tell everything from the capacity of the wallet.