Different things for different purposes. Measuring how real software under real production workloads in variable enviornments behaves is useful but inherently h

Different things for different purposes. Measuring how real software under real production workloads in variable enviornments behaves is useful but inherently high-variance. It doesn't let you track <1% changes commit-by-commit. Field work vs. lab work.