Firefox 72 vs. Chrome 80 Browser Performance On Ubuntu Linux With AMD Ryzen

[treba]

I still wonder if there is some build of Firefox with V8 rather than SpiderMonkey. Such a build would go a long way towards isolating performance problems - at least with regards to JavaScript related ones.

Well, some of the benchmarks here are very synthetic. Spidermonkey is much more optimized for real work websites (i.e. lots of "cold" unused js) than for "applications you better do in c". Because if you do want a heavily optimized application in the browser, you better use wasm - and that's where firefox outperforms chrome.
So the question would be if spidermonkey performs worse e.g. on the top 100 websites. And the chances are high that it actually performs much better.

See for example https://www.reddit.com/r/firefox/com...key_makes_the/

[andyprough]

Now do Brave browser.

[sunnyflunk]

Is this using the firefox from the Ubuntu repository or the binary provided from Mozilla?

I'm not sure the Ubuntu package is built with PGO so their version can be a significant improvement.

[kozman]

For interested parties, here are some current developments in Firefox to be exited about:
- experimental DMABUF support for WebGL (should bring it on par with Chrome) (ff74 https://bugzilla.mozilla.org/show_bug.cgi?id=1586696)
- experimental HW video decoding (74/75 https://bugzilla.mozilla.org/show_bug.cgi?id=1610199)
- work on partial damage (required for webrender rollout on battery powered devices) https://bugzilla.mozilla.org/show_bug.cgi?id=1484812
- as the above things all rely on EGL:
- ongoing work to make the EGL backend support OGL (not only GLES) https://bugzilla.mozilla.org/show_bug.cgi?id=1474281
- use the EGL backend on X11, too (currently only on Wayland) https://bugzilla.mozilla.org/show_bug.cgi?id=788319

So expect FF to make a huge jump in 2020

It's not quite fair to battle FF 72 against Chrome 80 with 73 just at RC1 as of 2 days ago. 73 isn't going to be a huge perf bump over 72 but 74 likely will be. But yes, lots of exciting stuff coming for FF in 2020. A lot of the Rust stuff, if they can square it away faster, should put it on par with Chrome in the not too distant future.

[xnor]

It's not quite fair to battle FF 72 against Chrome 80 with 73 just at RC1 as of 2 days ago. 73 isn't going to be a huge perf bump over 72 but 74 likely will be.

What are you talking about? It absolutely is fair. To suggest that a benchmark against an unreleased and unfinished version would be more fair is delusional.
FF is slower in many benchmarks. Just deal with it.

PS: I'm a FF user. Have tried Chromium several times and always went back.

[kozman]

What are you talking about? It absolutely is fair. To suggest that a benchmark against an unreleased and unfinished version would be more fair is delusional.
FF is slower in many benchmarks. Just deal with it.

PS: I'm a FF user. Have tried Chromium several times and always went back.

Delusional? What version of Chrome was out when FF 72 was out? Compare the version of Chrome that was out when FF 72 was released. That *would* be fair and yield some semblance of parity. And if Chrome 81 came out tomorrow? That would still be fair against FF 72? Chrome's cadence is far faster than FF's. Go ahead though, give all your privacy away to Google. No one gives a rat's posterior but you. Just deal with it. In real world use, the average user doesn't give a squirt what browser they're using.

[NateHubbard]

Delusional? What version of Chrome was out when FF 72 was out? Compare the version of Chrome that was out when FF 72 was released. That *would* be fair and yield some semblance of parity. And if Chrome 81 came out tomorrow? That would still be fair against FF 72? Chrome's cadence is far faster than FF's. Go ahead though, give all your privacy away to Google. No one gives a rat's posterior but you. Just deal with it. In real world use, the average user doesn't give a squirt what browser they're using.

Pretty sure it isn't going to make any real difference when you're talking about +/- a version number or two.
Anyway, I don't browse benchmarks, so as long as they are reasonably close and I don't notice any issues, I'm not going to switch from Firefox.

[marlock]

I wonder if any part of the performance gap between Chromium and Firefox is similar to the Intel vs. AMD CPUs performance gap of the recent past...

Anyone familiar enough with Chromium to comment on security and privacy shortcuts they may have taken that would grant a performance advantage?

It's actually a question, not meant to spread FUD, and I realize the chance is much slimmer that security oversights may go unnoticed in Chromium because it's opensource unlike Intel's chip architecture and because browsers are one of THE targets for cybersecurity testing and peer review efforts.

The thing I expect may come up is more about privacy features that are in any way taxing on performance and yet Mozilla would opt to deploy by default, while Chromium would skip. But then again some privacy improvements may reduce processing cost.

And then again, there is definitely a point to be made about the performance difference between running benchmarks (which by design and definition *should* all go through any privacy protections unaltered) against real-life webpages (which have all those tracking, fingerprinting, bitcurrency mining, background-running javascripts running and taxing performance on Chromium much more than on Firefox... or is there a benchmark for that?

[dwagner]

All those optimizations are so irrelevant for browser performance in contrast to the one optimization that dwarfs them all: Using a decent ad blocker. Advertisements have been the one and only relevant source of perceivable browser-slowness for so long now...

[xnor]

Delusional? What version of Chrome was out when FF 72 was out? Compare the version of Chrome that was out when FF 72 was released. That *would* be fair and yield some semblance of parity. And if Chrome 81 came out tomorrow? That would still be fair against FF 72? Chrome's cadence is far faster than FF's. Go ahead though, give all your privacy away to Google. No one gives a rat's posterior but you. Just deal with it. In real world use, the average user doesn't give a squirt what browser they're using.

Yeah, delusional to the point that you didn't even read what I wrote. Like the fact that I'm not using Chrome but FF.

Also, to suggest that a comparison today should be done with an older Chrome version when FF did a release is utterly ridiculous. The performance you get today is the performance the latest versions provide.
And if Chrome 81 came out tomorrow? Then tomorrow Michael would have to redo the benchmark for us to see what performance users with get with the latest versions; what performance users would get tomorrow. Doh!
"Fairness" has nothing to do with it which is why I called your comment delusional in the first place.

No one gives a rat's posterior but you. Just deal with it. In real world use, the average user doesn't give a squirt what browser they're using.

Wow, why so salty? It was you who brought up the ridiculous point of "fairness" in comparisons by comparing an outdated Chrome version to latest FF.
It was also you who just wrote "go ahead though, give all your privacy away to Google".

That to me sounds like someone giving a rat's ass. So you've just called yourself a no one, haven't you?
You really should calm down a bit...