Skip to content
This repository has been archived by the owner on Sep 30, 2023. It is now read-only.

Comparing baseline benchmarks #26

Open
mistakia opened this issue Jan 20, 2021 · 3 comments · May be fixed by #29
Open

Comparing baseline benchmarks #26

mistakia opened this issue Jan 20, 2021 · 3 comments · May be fixed by #29

Comments

@mistakia
Copy link
Contributor

mistakia commented Jan 20, 2021

To make it easier to compare benchmarks, I was thinking about adding a cli flag that would be used to write the baseline benchmark results to a provided path. Then this file can be loaded to provide a diff between two baseline benchmarks (i.e. comparing a PR to master).

Let me know if there's a better way to do this and how it could possibly integrate with CircleCI so that perhaps we could automate comparing baseline benchmarks for PRs.

@aphelionz
Copy link
Contributor

I think this sounds like a good idea as any. Anything that would get us automated comparative benchmarks like this is welcome.

@mistakia
Copy link
Contributor Author

mistakia commented Jan 20, 2021

My understanding of CircleCI is very limited but I was imagining it would run the benchmark in master and provide a globally accessible path. Then it would run the benchmark in that branch and the benchmark runner has access to that previous path and can output a report.

Not sure if this is possible.

@aphelionz
Copy link
Contributor

Yes, I believe it is via build artifacts. If you find it's easier in a different CI system i.e. GH actions that's fine too. I like CircleCI because I like a healthy ecosystem of tools instead of getting further locked-in to GH, but we're kinda there already anyway :P

@mistakia mistakia linked a pull request Feb 2, 2021 that will close this issue
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants