Closed Bug 1646694 Opened 4 years ago Closed 4 years ago

3.43 - 3.5% raptor-sunspider-firefox (windows10-64-shippable, windows10-64-shippable-qr) regression on push 58ad1a0b84193ce0e06d440741865c021d2fe0cc (Fri June 12 2020)

Categories

(Core :: JavaScript: GC, defect)

79 Branch
defect

Tracking

()

RESOLVED WONTFIX
Tracking Status
firefox-esr68 --- unaffected
firefox-esr78 --- unaffected
firefox77 --- unaffected
firefox78 --- unaffected
firefox79 - wontfix

People

(Reporter: Bebe, Unassigned)

References

(Regression)

Details

(Keywords: perf, perf-alert, regression)

Perfherder has detected a raptor performance regression from push 58ad1a0b84193ce0e06d440741865c021d2fe0cc. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

4% raptor-sunspider-firefox windows10-64-shippable-qr opt 173.62 -> 179.69
3% raptor-sunspider-firefox windows10-64-shippable opt 173.01 -> 179.00
3% raptor-sunspider-firefox windows10-64-shippable-qr opt 173.71 -> 179.67

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests. Please follow our guide to handling regression bugs and let us know your plans within 3 business days, or the offending patch(es) will be backed out in accordance with our regression policy.

For more information on performance sheriffing please see our FAQ.

Component: Performance → JavaScript: GC
Flags: needinfo?(sphink)
Product: Testing → Core

Set release status flags based on info from the regressing bug 1568923

Oops, I missed this, sorry! I could look at reducing the overhead, but I expect that it would end up being at the cost of less deduplication (eg if I just dropped the maximum length of strings eligible for deduplication.)

Given that this is Sunspider, I would say a 4% performance regression is an excellent result. Preventing pathological situations, as well as reducing memory usage somewhat in the common case, are well worth a 4% regression in Sunspider.

If this causes regressions in more important tests, I would want to look again.

(Note that I am not arguing for Sunspider to be removed from the set of monitored tests; I find this information useful.)

From https://www.mozilla.org/en-US/about/governance/policies/regressions/ it looks like I should just be arguing for WONTFIX, not marking it myself. Needinfo'ing :Bebe for that purpose.

Flags: needinfo?(sphink) → needinfo?(fstrugariu)

I'm OK with that resolution

Status: NEW → RESOLVED
Closed: 4 years ago
Flags: needinfo?(fstrugariu)
Resolution: --- → WONTFIX
Has Regression Range: --- → yes
You need to log in before you can comment on or make changes to this bug.