Mozilla's JM crosses Apple's Nitro threshold | Hacker News Reader

Mozilla's JM crosses Apple's Nitro threshold

arewefastyet.com

76 pointsdpurp36 comments

Open articleView on HN

A good browser benchmark would also measure:

1. Browser boot time.

2. Browser shutdown time (on clicking Close).

3. Tab boot time.

4. Tab shutdown time.

5. LocalStorage boot time (on first call to LocalStorage API, some browsers load LocalStorage into memory at this point, blocking the UI).

6. IO performance of LocalStorage, WebSQLDatabase, IndexedDB for massive get/set operations (bandwidth) and many get/set operations (latency).

7. ApplicationCache update time (and update time after an interrupted update cycle).

8. Smoothness of moving a DOM element across the screen by adjusting absolute position (some browsers are choppy with this).

9. Javascript parse time.

10. Minimum setTimeout frequency (is it 1ms or 5ms?).

11. Hash change event propagation time.

12. Stack size limit.

13. Speed of Array.shift on massive arrays.

14. Vertical height of browser chrome at top of the window in pixels (less is better).

15. Vertical height of browser chrome at bottom of the window in pixels (less is better).

16. Number of colors used in browser chrome, including favicons in location and bookmark bars etc. (less is better).

17. Fullscreen support and switch to fullscreen time.

18. Developer Console boot and shutdown time.

19. Number of words, tabs, nested interfaces used across all Preferences pages (less is better).

20. Number of user actions (clicks, keypresses, mouse movement distances) required to clear LocalStorage, History, WebSQLDatabase, IndexedDB, Cache, ApplicationCache (less is better).

21. Default storage limits for LocalStorage etc. (higher is better).

22. Data-URI size limits (higher is better).

Agreed that most of these would be good components of an overall browser benchmark, but the linked benchmarks are JavaScript benchmarks only, so things like browser start-up time and chrome dimensions wouldn't apply.

When users think "fast", they think "fast" as in "fast", not "JM fast" or "TM fast". It would be better for "arewefastyet.com" to address "fast" from the point of view of the user rather than from the point of view of the Javascript engine developer.

As I understand it, arewefastyet.com was created as more of an internal/community dashboard specifically for the JIT team at Mozilla rather than to reflect on the browser as a whole to the userbase.

I think 1 - 4 were what really kickstarted Chrome. Everyone else was focused on how many milliseconds difference between their rendering and js vs. the next guys, when Google shaved whole seconds off something that really matters.

Mozilla measures many of these at http://graphs.mozilla.org/. "Ts" is startup time; "Tshutdown" is shutdown time; "GFX" includes 8 I believe.

According to my friends that work at Mozilla, this site is purely for javascript speed and is what Mozilla uses as an internal motivator to motivate employees. Its a line everyone works to hit.

Its not really a "benchmark" to show the world that they're doing better. It's a public site because everything that happens in mozilla is open, but it's not really for us to see and judge. Its for the guys on the inside to see and cheer or see, then haul ass to catch up.

One possible takeaway from these graphs is the urgent need for better JavaScript benchmarks. Both SunSpider and the V8 benchmarks are sets of microbenchmarks and the difference between them is enormous: JM+TM is significantly faster than plain JM on V8 but is a little slower on SunSpider. The fact that the conclusions can vary so greatly among these two (both deeply flawed) benchmark suites suggests that the true "real-world" performance of the engines is all but entirely unmeasured. Which loads Gmail fastest? That would be a step in the right direction.

Here's an alternative: http://krakenbenchmark.mozilla.com/kraken-1.0/driver.html

I have to agree, that things like machine translated scheme and OS scheduling are not representative of real world applications. (Part of the V8 suite: http://v8.googlecode.com/svn/data/benchmarks/v6/run.html - specifically Richards, and EarleyBoyer)

Keep in mind that V8 is not only running in the browser, but has completely different usecases as well (see node.js).

That it's a javascript benchmark only means that.. it's used by google to finetune their code.. not created to compare javascript in browsers among several different browsers.

And yet that's what people seem to like using it for.

I still don't see why machine translated scheme is a good benchmark for fine tuning code. Javascript is a fine language to work in - especially if you are wanting to use it for things other than a browser (then you must really like it!).

I haven't seen the code but based on the description it's a benchmark that was originally in scheme and ported to javascript.. sounds fine to me. And "kernel simulation benchmark" may have a usecase too, just because you can't think of it, doesn't mean it isn't there ;) I can think of tests for scheduling, threading, asynchronous code, I/O when i think of "kernel simulation".

Again, people like to use V8benchmarks, when it's not made for browser comparisons. The fact that it's often used only shows that there is a lack of a real cross browser benchmar suite.

"...Florian Loitsch's Scheme2Js compiler"

No, Florian Loitsch isn't a human translator for hire - he wrote a compiler.

Yes, but the benchmark is not "how fast does V8 translate scheme to JS", it's "how fast does V8 run benchmarks in javascript that where translated from scheme some time ago". Atleast that's how i understand it.

That's true, but machine-generated code can be pretty nutty, depending on what exactly the Scheme2JS compiler does. It's legitimate JS code, but it might not be representative of what JS-written-in-JS looks like.

True, but i see that this is a valid testcase. For example GWT generates JS out of Java, so it might be useful to also benchmark machinegenerated JS.

I think it'd be useful to benchmark GWT-generated JS (and also Objective-J), but not Scheme2Js, which is rarely if ever used in practice. Compilers for different languages will emit code with completely different performance characteristics and feature usage patterns. (For instance, Objective-J particularly stresses the call site caching mechanisms in modern JavaScript engines with its "objc_msgSend" equivalent, since the indirect call in that function branches to so many different call sites.)

If you create a benchmark that measures X, it will be used to measure Y. It's the rule of benchmarks.

the richards benchmark is not only a simulation of OS scheduling though, it's a classic benchmark of method dispatch performance (at least from what I recall, which is why everybody is using it via a Smalltalk inheritance and not reimplementing the bcpl code)

About a month ago, Mozilla released their' "Kraken" benchmark suite, designed to better reflect what actual web-apps actually do:

    http://blog.mozilla.com/rob-sayre/2010/09/14/release-the-kraken/

It's not listed on arewefastyet.com, but apparently Firefox 4 handily beats Chrome and IE9:

    http://weblogs.mozillazine.org/asa/archives/2010/09/javascript_performan.html

The tests are: astar, beat-detection, dft, fft, oscillator, gaussian-blur, darkroom, desaturate, parse-financial, stringify-tinderbox, crypto-aes, crypto-ccm, crypto-pbkdf2, crypto-sha256-iterative.

One big thing that's missing here is DOM interaction, which is a real bottleneck for current web apps (and handling 'foreign' API well is JS engine's job too).

This appears to be what got them there: http://andreasgal.wordpress.com/2010/10/13/compartments/

Also worth noting, there is no longer a gigantic "NO" down at the bottom of the page. :)

I miss that "NO".

I expected to see "YES!", or maybe "MAYBE" in its place, but there's nothing there.

Yeah I was hoping for MAYBE, and then if it ever gets quicker than chrome a YES!

What actually surprises me about this is just how much faster V8 was (and really, still is) compared to all of the others. It also looks like they're starting to hit a point of diminishing returns. Here's hoping they're not!

Only x86, though. x64 isn't there yet. http://arewefastyet.com/?machine=4

I am always puzzled by the attention paid to JavaScript execution speed. Obviously, it's important (and should be) to the people who are intimately involved with the maintenance and development of the JavaScript engines, but apart from that, why do so many people in the web business care about it so much?

What sites and/or web apps are so JavaScript-heavy that they are bottlenecked by slow JavaScript execution?

In most cases, aren't the speed of a user's network connection and the responsiveness of the web server itself far more important to the user experience of a web app than the browser's JavaScript speed? Is GMail on Chrome really a radically different experience from GMail on IE8?

People are looking ahead to html 5, when real apps can be implemented entirely in a browser.

I have heard from the V8 developers that Mozilla do some optimizations just to score better on benchmarks - i.e. they score really good on things that are specifically benchmarked. I have not verified their claim, but generally for me as an user Chrome and Safari feel a lot faster than Firefox.

Two quick points (I work at Mozilla, but not on the browser - I'm also the creator of the Dromaeo performance benchmark): All browsers implement things to get faster at these benchmarks. They implement caching in bizarre places just to make an, otherwise, useless loop faster. They optimize methods that are rarely used in modern JavaScript applications. This frustrates me immensely. Most modern browser performance benchmarks focus exclusively on JavaScript and completely ignore the DOM, rendering, painting, or any of the other operations that are critical to the performance of a web site.

As to the faster claim - I've been using Firefox 3.6, Chrome, Safari 5, and Firefox 4 interchangeably and I've noticed a dramatic improvement in quality from Firefox 3.6 to Firefox 4 - they've really done a great job with the upcoming release.

Especially Sunspider, which has the doubly whammy of developer and to a certain extent power-user mindshare, but yet doesn’t test normal site behaviour.

Sunspider is a particularly bad benchmark: Virtually all of the components are heavily weighted by string concatenation.

http://www.yafla.com/dforbes/String_Concatenation_and_Immuta...

Agreed. I'm just using development builds (and lately the official beta), but Firefox 4 is leaps and bounds better than Firefox 3.6 (which was already pretty good!)

Have you had problems where Firefox 4 just fails miserably to render a page, but if you refresh it renders fine?

I bet slow JM has cost FireFox a pretty significant number of users over the last couple of years. Even the much smaller gap between nitro & v8 is noticeable at times.

You have no idea what you're talking about. If you recognize it as speculation, please don't do it when actual verification is such a simple process.

Pre-emptive note to readers: Sure "You have no idea what you're talking about" sounds harsh, but is it wrong? Please consider that before acting upon any knee-jerk reactions to my comment.

Reply on news.ycombinator.com