Bug 1849393 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.

I have reliable STR to trigger this bug (starting with a fresh profile) over in bug 1867137 comment 0 --  "just" create a window with 500 https://bugzilla.mozilla.org/ tabs (quick steps to do this over in the other bug).  (Bugzilla is particularly useful for this since it's favicon is kinda heavyweight, and it gets encoded into the sessionstore for each tab as a data URI, per comment 11.)

With that window open in the background, Firefox seems to just repeatedly jank for 500ms and undergo a 225MB jump in memory usage, every 15s. (Presumably the memory usage is freed periodically once it hits some GC threshold, and that causes another period of jank.)

Here's a profile: https://share.firefox.dev/41fFP5D
I have reliable STR to trigger this bug (starting with a fresh profile) over in bug 1867137 comment 0 --  "just" create a window with 500 https://bugzilla.mozilla.org/ tabs (quick steps to do this over in the other bug).  (Bugzilla is particularly useful for this since its favicon is kinda heavyweight, and it gets encoded into the sessionstore for each tab as a data URI, per comment 11.)

With that window open in the background, Firefox seems to just repeatedly jank for 500ms and undergo a 225MB jump in memory usage, every 15s. (Presumably the memory usage is freed periodically once it hits some GC threshold, and that causes another period of jank.)

Here's a profile: https://share.firefox.dev/41fFP5D
I have reliable STR to trigger this bug (starting with a fresh profile) over in bug 1867137 comment 0 --  "just" create a window with 500 https://bugzilla.mozilla.org/ tabs (quick steps to do this over in the other bug).  (Bugzilla is particularly useful for this since its favicon is kinda heavyweight, and it gets encoded into the sessionstore for each tab as a data URI, per comment 11.)

With that window open in the background, Firefox seems to just repeatedly jank for 500ms and undergo a persistent 225MB increase in parent-process memory usage, every 15s. (Presumably the memory increase resets periodically once it hits some GC threshold, and that massive multi-gigabyte GC presumably causes another period of jank.)

Here's a profile: https://share.firefox.dev/41fFP5D
I have reliable STR to trigger this bug (starting with a fresh profile) over in bug 1867161 comment 0 --  "just" create a window with 500 https://bugzilla.mozilla.org/ tabs (quick steps to do this over in the other bug).  (Bugzilla is particularly useful for this since its favicon is kinda heavyweight, and it gets encoded into the sessionstore for each tab as a data URI, per comment 11.)

With that window open in the background, Firefox seems to just repeatedly jank for 500ms and undergo a persistent 225MB increase in parent-process memory usage, every 15s. (Presumably the memory increase resets periodically once it hits some GC threshold, and that massive multi-gigabyte GC presumably causes another period of jank.)

Here's a profile: https://share.firefox.dev/41fFP5D
I have reliable STR to trigger this bug (starting with a fresh profile) over in bug 1867161 comment 0 --  "just" create a window with 500 https://bugzilla.mozilla.org/ tabs (quick steps to do this over in the other bug).  (Bugzilla is particularly useful for this since its favicon is kinda heavyweight, and it gets encoded into the sessionstore for each tab as a data URI, per comment 11.)

With that window open in the background, Firefox seems to just repeatedly jank for 500ms and undergo a persistent ~225MB increase in parent-process memory usage, every 15s. (Presumably the memory increase resets periodically once it hits some GC threshold, and that massive multi-gigabyte GC presumably causes another period of jank.)

Here's a profile: https://share.firefox.dev/41fFP5D

Back to Bug 1849393 Comment 14