Loading this sheet and waiting for the graphs to stop updating, it takes about 3.9 seconds for me in Firefox and 2.6 seconds in Chrome.

https://docs.google.com/spreadsheets/d/1jfSqXZGjIVl5FJoSxoElMrXeIFJKOmmqgere1v3PddQ/edit#gid=0

Taking a quick glance at a profile, nothing immediately stands out. Would your profile-comparison tooling be able to identify places where we're slower?

(In reply to Iain Ireland [:iain] from comment #1)

Would your profile-comparison tooling be able to identify places where we're slower?

Yup, working on it! Preliminary report is here: https://github.com/jrmuizel/js-profile-compare/blob/main/reports/Google-Sheets-browser.md
It might have some misattributed samples; I'll generate a new report tomorrow that should have those fixed.

Oh, sweet!

What are you doing for inlining here? I notice that some of the hot functions are quite small, which makes me wonder if they're being inlined in one browser.

We've generated the report with better profile data. It should be accurate now!

(In reply to Iain Ireland [:iain] from comment #3)

What are you doing for inlining here?

We're turning it off, specifically we're running with --no-turbo-inlining in Chrome and --only-inline-selfhosted in Firefox. (Well, for Firefox, we changed the code because we didn't know how to pass this argument when using Firefox instead of the shell.)

So the inlining should be the same in the two browsers. The original report had some addresses resolved to incorrect JS function names, which may have confused things.

The Performance Impact Calculator has determined this bug's performance impact to be medium. If you'd like to request re-triage, you can reset the Performance Impact flag to "?" or needinfo the triage sheriff.

Platforms: Windows
Impact on browser: Causes noticeable jank
Impact on site: Causes noticeable jank
Page load impact: Some

Report from Windows: https://github.com/jrmuizel/js-profile-compare/blob/main/reports/Google-Sheets-BrowserWin-May17.md