Benchmarking Firefox 83 Nightly With "Warp" Against Google Chrome On Linux

[Jumbotron]

One man's progress is the other man's regress 😋

Mozilla ought to name the next version of Firefox...."Tenet"

[oleid]

Maybe I'm misreading the graphs, but it looks like 83 is generally a big step backwards and Warp didn't make much of a difference in these benchmarks.

Well, there is an regression, as mentioned in the text. And it is a nightly build with lots of debug stuff enabled.

[Brane215]

I don't care about few 10% here and there, as long as the browser is lean.
Insane unlooping and similar approaches just to capture that last 5% of friggin JS engine seem insasne waste of time while there obviously are ways to circumvent whole VM protection layer.
Mozilla's problem is virtue signaling in an entity that was supposed to be about a product within given social contract.
So, had they focused on new approaches ( kudos for Rust ! - rare exemption), like having a browser as a local native container for running remote native programs instead of JS or something similar, even as a testing project, they would have much more market attention...

[oleid]

having a browser as a local native container for running remote native programs instead of JS or something similar,

You mean like webassembly? Or what exactly do you have in mind?

[ferry]

I think we all already knew Chrome is faster, and FF uses less memory per tab. Unfortunately, the latter was not tested afaikt.

Would the complainers be happy when FF would have beat Chrome performance at the expense of doubled memory use?

BTW I have no problems with web site compatibility to FF. A while ago I did have repeatable problems with Google Earth on Chromium freezing linux (hard reset required).

[pal666]

i wonder when ff will follow suit and wrap chromium like everyone else

[pal666]

BTW I have no problems with web site compatibility to FF.

but i do

[crystall]

Maybe I'm misreading the graphs, but it looks like 83 is generally a big step backwards and Warp didn't make much of a difference in these benchmarks.

You can't really compare Firefox nightly performance to the release version. Nightly has a number of debugging / telemetry features enabled that aren't in release builds.

[mphuZ]

i wonder when ff will follow suit and wrap chromium like everyone else

After 2022.

[caligula]

Well, there is an regression, as mentioned in the text. And it is a nightly build with lots of debug stuff enabled.

Um, so you're saying that the way modern agile software development works is that for every development branch, they first add shitloads of debug stuff that slows everything down, without adding any options for turning it off? Then, at the end of the release process, they quickly remove all the debug stuff and the bugs, and compile a shiny slim production version that nobody could build from the intermediate states? Sounds like a plan.. not. The thing is, there's simply no need to break absolutely everything before every new release.