Bug 1588710 Comment 14 Edit History

Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.

Dave told me the following:
>  If so, this suggests that your patch has caused a high confidence regression to the kraken results across all platforms. The kraken benchmark measures JavaScript performance, details can be found at https://wiki.mozilla.org/TestEngineering/Performance/Talos/Tests#kraken. There also appears to be a high confidence improvement to tabpaint across all platforms, which is documented here: https://wiki.mozilla.org/TestEngineering/Performance/Talos/Tests#tabpaint. There are a few other medium to high confidence improvements and regressions.

> I see references to linux64, macos64, and win64 in your patch here https://hg.mozilla.org/try/rev/408e7270af623961e8cb466aa31b18248aa3e512 could these explain the cross-platform impact? Were the tasks executed around the same time? The reason I ask is that if the base tasks were triggered first, and the new triggered much later, perhaps an infrastructure change could have been introduced in the meantime.

Serge told me that he will have a look why Windows performances are impacted
Dave Hunt told me the following:
>  If so, this suggests that your patch has caused a high confidence regression to the kraken results across all platforms. The kraken benchmark measures JavaScript performance, details can be found at https://wiki.mozilla.org/TestEngineering/Performance/Talos/Tests#kraken. There also appears to be a high confidence improvement to tabpaint across all platforms, which is documented here: https://wiki.mozilla.org/TestEngineering/Performance/Talos/Tests#tabpaint. There are a few other medium to high confidence improvements and regressions.

> I see references to linux64, macos64, and win64 in your patch here https://hg.mozilla.org/try/rev/408e7270af623961e8cb466aa31b18248aa3e512 could these explain the cross-platform impact? Were the tasks executed around the same time? The reason I ask is that if the base tasks were triggered first, and the new triggered much later, perhaps an infrastructure change could have been introduced in the meantime.

Serge told me that he will have a look why Windows performances are impacted

Back to Bug 1588710 Comment 14