Benchmark using cachegrind #1137
Conversation
FedericoCeratto
force-pushed
the
cachegrind
branch
from
ce391d0 to
cc58e80
Compare
FedericoCeratto
force-pushed
the
cachegrind
branch
from
cc58e80 to
bffdceb
Compare
| @@ -1,6 +1,7 @@ | |||
| # Runs quality assurance checks | |||
| name: "qa" | |||
| on: | |||
| pull_request: | |||
There was a problem hiding this comment.
We don't want to run this workflow on every pull request because it is slow and it will hence slow down development. We usually only want to run QA tests when we are inside a release branch.
There was a problem hiding this comment.
I don't think we should conflate the scripts we use for performing QA with cachegrind runs because that would most likely make the QA scripts (which are already slow and flaky) even slower and more annoying to run during releases.
Go contains an extensive framework for benchmarking, but it is unclear to me whether there is an equivalent of using cachegrind. Could you check? (I'd rather really use native Go tools for benchmarking than external tools.)
What were you trying to benchmark? Perhaps, I can help by adding support to miniooni for generating performance traces that we can analyze with Go tools.
Implement basic reproducible benchmarking
See https://pythonspeed.com/articles/consistent-benchmarking-in-ci/