Javascript Performance Thread

Discussion about official Mozilla Firefox builds
Post Reply
neo86
Posts: 124
Joined: July 21st, 2008, 2:53 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by neo86 »

bzbarsky wrote:
neo86 wrote:Hmm, I was hoping for something more exciting. :neutral:

More exciting than which? And what excites you? ;)


Well... both - something along the lines of the JSNES emulator. Even in comparison to recent JM builds Chrome was faster, but I think there were some code changes in the emulator which has lowered the performance in both browsers. As for the second question that's between me and the ladies. ;-)
User avatar
bzbarsky
Posts: 478
Joined: November 5th, 2002, 1:36 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by bzbarsky »

Tom B. wrote:Hmm, I'm surprised by the replies. For me, Google Chrome is so much faster. GC renders the table in the blink of an eye, while TM is very slow, and I can feel the CPU lag

Just to make sure... is this in a clean profile with no extensions and the jit (both of them, in the case of TM) enabled?

It's possible that the bottlenecks are different on different hardware here, though....
User avatar
bzbarsky
Posts: 478
Joined: November 5th, 2002, 1:36 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by bzbarsky »

Cru_N_cher wrote:http://js1k.com/demo/653 (almost freezes the GUI, shows nicely how heavily firefox needs seperate GUI processing like GChrome has it)

In an m-c build, yes. In a JM/TM build, not here....

http://29a.ch/2010/6/2/realtime-raytracing-in-javascript

Yeah, there's serious memory thrashing going on here. See analysis in https://bugzilla.mozilla.org/show_bug.cgi?id=590379 from a few weeks ago.

http://vis.stanford.edu/protovis/ex/force.html

Again, faster in JM/TM than on m-c but needs more work. See https://bugzilla.mozilla.org/show_bug.cgi?id=589422 for the gory details, but the upshot is that about half the time here is JavaScript and the other half is other stuff; SVG path performance shows up prominently.

and it goes on

http://acko.net/blog/js1k-demo-the-making-of

That looks about the same to me in Chrome 7 dev and in JM/TM.

No one is saying that there's nothing else to do, of course. But at this point there are also plenty of examples where we're faster than the competition; now all we need to do is be faster on all of them. ;)
User avatar
bzbarsky
Posts: 478
Joined: November 5th, 2002, 1:36 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by bzbarsky »

neo86 wrote:something along the lines of the JSNES emulator.

I guess I'm more excited by useful tools like the lightroom thing than by NES emulators....
Tom B.
Posts: 157
Joined: May 3rd, 2004, 3:53 am
Location: London, England

Re: Firefox 4.0 TM and JM Performance Thread

Post by Tom B. »

bzbarsky wrote:
Tom B. wrote:Hmm, I'm surprised by the replies. For me, Google Chrome is so much faster. GC renders the table in the blink of an eye, while TM is very slow, and I can feel the CPU lag

Just to make sure... is this in a clean profile with no extensions and the jit (both of them, in the case of TM) enabled?

Yes — I used a brand new profile, and I ensured that everything in about:config containing the string "jit" was enabled.
neo86
Posts: 124
Joined: July 21st, 2008, 2:53 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by neo86 »

bzbarsky wrote:
neo86 wrote:something along the lines of the JSNES emulator.

I guess I'm more excited by useful tools like the lightroom thing than by NES emulators....


Yea... I'm sure the general public, of which I'm a part of, would agree with you rather than me. Even if you think that NES emulator is stupid it's much more entertaining. Entertainment means ad revenue. That "lightroom thing" (also referred to as Darkroom) was terribly slow regardless of which browser you used. Yes... JM build was 3x faster than Chrome, but 10 fps is still pretty unusable - not particularly useful. It would take less time to just install Photoshop or Paint.NET. JSNES, even though it's slower now, is still a much more enjoyable experience rather than a frustratingly slow one.
User avatar
bzbarsky
Posts: 478
Joined: November 5th, 2002, 1:36 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by bzbarsky »

neo86 wrote:Even if you think that NES emulator is stupid it's much more entertaining.

Sure. I just think the other is much more indicative of the way the web can develop with fast JS engines; while games are certainly important (both on the web and off), so are other things.

That "lightroom thing" (also referred to as Darkroom) was terribly slow regardless of which browser you used. Yes... JM build was 3x faster than Chrome, but 10 fps is still pretty unusable - not particularly useful.

I'm seeing close to 20fps here, but that may be a hardware issue. And for what it's doing, which is not realtime video, even 10fps is pretty usable. The key part there is that there be no perceptible lag between moving the slider and the image changing; was there for you at 10fps?

In any case, I'm more interested in moving the web forward than in marketing here, which affects what I get excited about.
wellofsouls
Posts: 168
Joined: March 22nd, 2008, 10:12 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by wellofsouls »

bzbarsky wrote:In an m-c build, yes. In a JM/TM build, not here....

...

That looks about the same to me in Chrome 7 dev and in JM/TM.

No one is saying that there's nothing else to do, of course. But at this point there are also plenty of examples where we're faster than the competition; now all we need to do is be faster on all of them. ;)

I'm not sure if people are talking about the same thing, are we comparing Chrome to the current Firefox, or JM/TM Firefox?

I think it started with something about a Chrome spin from Webmonkey, and when I read the Webmonkey article, it seems to just say that Chrome is faster than the current Firefox, while the JM/TM Firefox is as fast as Chrome.

So I'd guess the "name a JS program where Firefox is faster than Chrome" thing, when talking about Webmonkey's "Chrome spin", is about the current Firefox and Chrome. And I'd think saying "the current/m-c Firefox is slower than Chrome, the JM/TM Firefox is as fast as Chrome" is exactly what Webmonkey's "Chrome spin" is saying.

Just my two cents. Maybe I have missed something, oh well...
Unbiased (as much as possible at least :p ) Opera Fan
Opera snapshot / Konqueror trunk / Firefox Nightly / Chrome Canary / IE10
Windows 7 / Windows 8 / Arch Linux / Mac OS X
User avatar
malliz
Folder@Home
Posts: 43796
Joined: December 7th, 2002, 4:34 am
Location: Australia

Re: Firefox 4.0 TM and JM Performance Thread

Post by malliz »

Like the fact bzbarsky is a Mozilla Dev? Give it the low level flaming a rest for a while ok wellofsouls ? It is getting really old
What sort of man would put a known criminal in charge of a major branch of government? Apart from, say, the average voter.
"Terry Pratchett"
User avatar
bzbarsky
Posts: 478
Joined: November 5th, 2002, 1:36 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by bzbarsky »

wellofsouls wrote:So I'd guess the "name a JS program where Firefox is faster than Chrome" thing, when talking about Webmonkey's "Chrome spin", is about the current Firefox and Chrome.

Ah. I think given the context of the rest of this discussion, that was very unclear. Firefox 3.6 is very old news. ;)
wellofsouls
Posts: 168
Joined: March 22nd, 2008, 10:12 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by wellofsouls »

malliz wrote:Like the fact bzbarsky is a Mozilla Dev? Give it the low level flaming a rest for a while ok wellofsouls ? It is getting really old

well, first, I didn't know bzbarsky is a Mozilla Dev, so thanks for pointing that out, but I don't know how that's related to the discussion here.

second, I don't know what "low level flaming" you are talking about, so I'd appreciate it if you can tell me who/what did I flame with that post of mine? I'm just pointing out the fact that this whole thing seem to have started with some mention of "Webmonkey's Chrome spin", back with this very post
neo86 wrote:
Omega X wrote:
Erunno wrote:<quote>The "press" also got word of it.

Webmonkey as usual took the typical Chrome spin.<quote>

Not everyone prefers to put on the blinders when it comes to the competition. :wink:


Says, the chief of Chrome spin.


Can you name a JS program where Firefox is faster than Chrome?

And I don't see how pointing out that fact could have possibly offended you, or anyone else, or what's the flaming and what's getting old? :?:
Unbiased (as much as possible at least :p ) Opera Fan
Opera snapshot / Konqueror trunk / Firefox Nightly / Chrome Canary / IE10
Windows 7 / Windows 8 / Arch Linux / Mac OS X
wellofsouls
Posts: 168
Joined: March 22nd, 2008, 10:12 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by wellofsouls »

bzbarsky wrote:
wellofsouls wrote:So I'd guess the "name a JS program where Firefox is faster than Chrome" thing, when talking about Webmonkey's "Chrome spin", is about the current Firefox and Chrome.

Ah. I think given the context of the rest of this discussion, that was very unclear. Firefox 3.6 is very old news. ;)

yea, I just think it's safe to say that Firefox with JM/TM is/will be among the fastest out there, if not *the* fastest, and everyone should agree with that.

oh, and keep the great work up, you guys rock 8-)
Unbiased (as much as possible at least :p ) Opera Fan
Opera snapshot / Konqueror trunk / Firefox Nightly / Chrome Canary / IE10
Windows 7 / Windows 8 / Arch Linux / Mac OS X
Cru_N_cher
Posts: 377
Joined: January 31st, 2010, 11:15 am

Re: Firefox 4.0 TM and JM Performance Thread

Post by Cru_N_cher »

@bzbarsky
is JM/TM speeding up any part of Tab Candy (Firefox Panorama) or the Extensions (Chrome) themselves yet ?, that would be i guess useful (combined with the GPU acceleration), especially if you look @ the future of it (plugins) different visualizations (more heavy load) ;)
WonderCsabo
Posts: 2230
Joined: June 25th, 2010, 9:29 am
Location: Budapest, Hungary

Re: Firefox 4.0 TM and JM Performance Thread

Post by WonderCsabo »

Of course, these things are written in JS.
Dorus
Posts: 238
Joined: September 6th, 2004, 2:18 pm

Re: Firefox 4.0 TM and JM Performance Thread

Post by Dorus »

WonderCsabo wrote:Of course, these things are written in JS.

Are you sure? I remember activating TM for chrome content was delayed quite a bit when it was new. Is this different for JM?
Post Reply