Thanks Diana a lot for offering to help. Please see my replies inline below.
On Sat, Sep 11, 2021 at 8:37 AM Diana Clarke <diana.joan.cla...@gmail.com> wrote: > If you point me to the existing benchmarks for each project and > instructions on how to execute them, I can let you know the easiest > integration path. > For arrow-datafusion, you just need to install the rust toolchain using `rustup` [1], then run the `cargo bench` command within the project root. Benchmark results will be saved under the `/target/criterion/BENCH_NAME/new` folder as raw.csv file. You can read more about the convention at https://bheisler.github.io/criterion.rs/book/user_guide/csv_output.html. For arrow-rs, it's the exact same setup. We have some extra TPCH integration benchmarks in datafusion, but I think we can work on integrating them later. Getting the basic criterion benchmarks into conbench would already be a huge win for us. [1]: https://rustup.rs > If the arrow-rs benchmarks are executable from command line and return > parsable results (like json), it should be pretty easy to publish the > results. > By default, results are saved as csv files, but you can pass in a `--message-format=json` argument to save the results as JSON files instead. > - The arrow-rs and arrow-datafusion GitHub repositories must use > squash merges (or Conbench would have to be extended to understand the > 2 other GitHub merge methods). Yes, we are using squash merge for both repos. > - I'm not sure what the security implications are with respect to > adding our ursabot integration and buildkite hooks to other > repositories. > Are you concerned about security from ursabot and buildkit's point of views? If so, who should we reach out to discuss this matter? Thanks, QP