-
Notifications
You must be signed in to change notification settings - Fork 29
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
feat: quick 3 run scheme #427
base: main
Are you sure you want to change the base?
Conversation
|
Performance Comparison ReportSignificant Changes To Duration
Show details
Meaningless Changes To DurationShow entries
Show details
Changes To CountThere are no entries Added ScenariosThere are no entries Removed ScenariosThere are no entries |
Summary
This experimental scheme of test runs will run the test 3 times (plus any additional warmup runs to be discarded). This will result in three test results: min, mid & max. We then modify the run table to contain the following content: [min, mid, mid, mid, mid, max], effectively quadrupling the middle measurement.
The idea is based on PERT distribution. Since I am not a statistician, I propose this PR as an engineering exercise when we would verify its usefulness against some stability test. The main benefit being in reducing the default 11 runs (10 real + 1 warmup) to just 4 runs (3 runs + 1 warmpu).
Test plan
This method should undergo stability testing to compare how well or badly it fares against the default
run: 10
scheme.