Open Bug 1258249 Opened 4 years ago Updated 4 years ago

Windows debug reftests take almost 3x longer to run than opt

Categories

(Testing :: Reftest, defect)

Version 3
Unspecified
Windows
defect
Not set

Tracking

(Not tracked)

People

(Reporter: RyanVM, Unassigned)

Details

That seems like a crazy-huge discrepancy to me, especially on dedicated hardware. And it can't be good for wait times. I don't remember it always being this bad, but a cursory glance at the mozilla-release repo says it goes back to at least Fx42. ESR38 shows closer to a ~2x slowdown (more in line with what we'd probably expect), so sometime after that I guess?

I know these bugs often fall into the "Not it!" category for investigation, but this feels like something that could yield positive results if someone takes the time to figure out why there's such a huge difference. And given our issues with Windows wait times, that time seems like it'd be well-spent.
OSX also shows a ~2x slowdown, so Windows is clearly the outlier here.
You need to log in before you can comment on or make changes to this bug.