Opera 9981 beats Firefox 3.0pre in a Mozilla benchmark


Recommended Posts

The Linux performance of Opera is still a bit on the slower side, but the Windows version just speeds up with every new snapshot release.

And now the newest (and fastest) snapshot of Opera (build 9981) actually beats the newest (and fastest AFAIK) nightly of Firefox (2008051011) in the Mozilla javascript performance test, the Dromaeo benchmark.

Opera score :

http://dromaeo.com/?id=9707

Firefox score :

http://dromaeo.com/?id=9709

The latest Opera snapshot :

http://my.opera.com/desktopteam/blog/2008/...d-optimizations

It's quite a feat when you consider that just more than a couple weeks ago Opera was behind Firefox by 1 whole second, and now they are ahead by 100ms, not to mention it's a test made by Mozilla :cool:

100 ms :| as if I care about it when I browse some web sites

well, it's a benchmark, and the total result is just around 2 second, so when you put it into perspective, from 1 second (50%) slower to 100ms (5%) faster in a couple of weeks is quite a feat ;)

Especially when Firefox made it sound like a huge thing by claiming itself to be the fastest browser when b5 was released, so it's fair game :p

Your result is in the 11 thousand range, the new Opera build is in the 2 thousand range ;)

Depends on the machine. My FFb5 is in the low 1000's http://dromaeo.com/?id=9815

Anywho, this is good. Faster browsers are better for us (the consumer). Yay for competition!

No doubt the latest Opera weekly is faster. However, my results tend to differ from yours:

Opera 9.5 beta build 9981: 1392

Minefield 2008051107: 1353

Firefox 2.0.0.14: 6992.6 (lol)

Webkit r33029: 1246.4

Results are too close to call for the latest Opera and Firefox previews. So let's just sit back and pat both on the back for speeding up their products. :yes:

Edit: Let's give Webkit some praise too.

Edited by rm20010
Ya in ms. big deal. Wait until your Opera becomes final. I have never heard of this test site. As far as I know it is meaningless.

well it's about as meaningless as how the Mozilla guys made it sound like a BIG DEAL when ff3b5 was released and claimed itself as the fastest browser for some synthetic javascript benchmark. And some dude here even posted a "Firefox beats everything" thread back then, so it's fair game :p

and obviously it's quite a big deal to them since they even made their own javascript benchmark, the Dromaeo, so yeah this test site is THE Mozilla javascript test, it's no wonder you haven't heard of it since Mozilla just made it after firefox3b5. If you have something against it, tell that to the Mozilla guys, and tell them to wait their Firefox becomes final instead of beta 5 :p

But then they both make some great progress in javascript performance lately, so kudos to both teams, and to the webkit for being very very fast all this time.

This is a generic browser test site that happened to have been made by a Mozilla (and jQuery) guy. Mozilla have taken it and are working with other browser guys to make the test better.

Not everything has to be a "us vs. them" thing.

This is a generic browser test site that happened to have been made by a Mozilla (and jQuery) guy. Mozilla have taken it and are working with other browser guys to make the test better.

Not everything has to be a "us vs. them" thing.

So? it doesn't change the fact that it is THE Mozilla javascript test. It's not a "us vs. them" thing, it's just a fair game after all those "Firefox 3b5 fastest browser" reports everywhere for the past month or so :p

I tried this in IE7 and it just froze and I had to end the process lol.

yea, IE7 freezes, IE8 beta 1 can pass the test, I guess that means the IE guys are also doing a good work lately (compared to themselves before that is) :p

I think the fact that when the flashing contest button on the NW frontpage was broken and used a full core on IE, and ~25%(50% of one core) on FF and what was it 5% or so on Opera shows wich browser handles scripts better :)

I still use Maxthon though, but I may change when 9.5 is final.

Ya in ms. big deal. Wait until your Opera becomes final. I have never heard of this test site. As far as I know it is meaningless.

Javascript performance isn't meaningless though, especially not in this day and age of Javascript-heavy AJAX sites.

So this is interesting, and I think there's no coincidence both Mozilla and Opera have been focusing a bit on this lately. Sites may widely be affected, regardless if it's about Gmail or e.g. Digg.

Well I did some benchmarks on my computer (x2 4800+ @2.64Ghz / WinXP) and it does seem to be speedy but Firefox 3.0pre 080512 (ayakawa's PGO build) had a slight edge over Opera 9981 in JavaScript when I tested. Overall both seem very fast and you probably wouldn't notice much of a difference between the two.

Opera Score 2278.20ms (Total)

http://dromaeo.com/?id=10000

Firefox Score 1987.60ms (Total)

http://dromaeo.com/?id=10004

JavaScript Performance Test

Firefox firefox30prewj4.png

Opera opera9981os3.png

DOM Core Performance Appending, Prepending, Inserting, Indexing, Removing

Firefox (avg. of 10 tests): 961.1ms

Opera (avg. of 10 tests): 511.1ms

DOM Core Performance Parsing

Firefox (avg. of 10 tests): 1359ms

Opera (avg. of 10 tests): 1467.7ms

hmm

if I add up the scores form the 3 tests you showed/listed I get

FF: 2585

Opera: 2330

Well if you add in the SunSpider Javascript benchmark in which Opera is about 2 seconds slower, then Firefox would take a considerable lead.

SunSpider Result Firefox 3770.4ms

SunSpider Result Opera 5775.4ms

All said and done optimized Firefox 3.0pre builds are still faster with javascript then Opera and Opera is still faster rendering the other objects on the page other than javascript. If they stay neck and neck like this while continuing to improve, performance won't even be much of a factor when making a decision to use either Opera 9.50 or Firefox 3.0 once the final versions are released.

well, it seems Firefox managed to improve its score steadily in the past few days with each new build

http://dromaeo.com/?id=10161

Also I guess it depends a lot on the system, maybe that's what the PGO is about. It's faster on the Woodcrest system, but a bit slower on my old Thinkpad

http://dromaeo.com/?id=10162

http://dromaeo.com/?id=10164

  • 2 weeks later...
  • 3 weeks later...
This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Posts

    • Seems to be working fine for me.
    • Latest Rufus update improves new Windows 11 install method by Taras Buria Pete Batard, the maker of Rufus, a very popular app for creating bootable Windows (and other OS) media, has released a new beta version of its app. Rufus 4.15 beta is now out, and while it offers no new features, there are all sorts of improvements and fixes, including for the new Windows 11 installation method that was introduced in version 4.14 in early May. The "Silent Windows 11 installation" is a new feature whose goal is to automate operating system installation. All you have to do is boot from the drive, and then Rufus takes over, doing all things for you, such as setting up a new account, skipping ads and prompts, and more. It is a very handy tool, but initially, it had some bugs and issues that required addressing. With version 4.15 beta, Rufus is fixing that, particularly a bug with installation failing at 75%, crashes on Snapdragon X-based PCs, and more. Here is the changelog: Rufus 4.15 beta is now available for download from its GitHub repository. If you have never used Rufus before, you can check out our guide here. It is a very useful utility to have, as it allows you to deal with plenty of Windows 11's annoyances, which are still there, despite Microsoft's ongoing efforts to fix them.
    • Microsoft fixes one of Excel Copilot's most frustrating limitations by Usama Jawad Microsoft began integrating Copilot into Excel a couple of years ago and has been upgrading it with new functionalities since then. While some changes have been controversial, Microsoft is hoping to win over users by allowing them to be more productive via Copilot. To that end, it has now announced a Copilot improvement that may actually be appreciated by people who use it regularly. Excel customers often use the Copilot prompt box to issue instructions to format and customize their data, but it can become quite tiring to keep repeating the same instructions again and again. Microsoft now allows you to define Copilot personalization rules for formatting, naming conventions, formulas, and report styles. These can be accessed via Settings > Personalization, where you can explain your rules in natural language like "Always format currency in USD with no decimals", and just let Copilot take care of the rest. Microsoft is going a step further in this direction by allowing you to set workbook rules too. These rules are stored as a .Rules sheet, and are preserved while the workbook is shared. This fosters collaboration while making sure that standard rules govern the Copilot editing experience across the organization. Other advantages of this capability include pointing it to specific examples, defining dynamic formulas, and referencing an entire sheet and asking Copilot to infer rules based on that. You can leverage this feature by opening Copilot in Excel, clicking on "+", and selecting Create workbook rules. If you have an existing .Rules sheet, you can simply start listing the rules in column A as well. Personalization features are available to all Copilot in Excel users across the web, Mac, and Windows. Meanwhile, workbook rules are currently being previewed for Windows and Mac customers on the Insiders channel. General availability is scheduled after a few weeks, but a concrete date is currently unknown. Overall, the Excel capability is quite similar to ChatGPT's memory features, which allow you to permanently store items in the AI model's context window.
    • Imagine you still haven't discovered Total Commander that is doing all those things for three decades already...
    • This sounds like underneath the nice marketing spin, either someone at Adobe got tired of their lazy devs and asked Microsoft to help them sort at least some of Adobe's ancestral spaghetti code to make it go faster, or Microsoft wanted Adobe's crap to run better on Windows to make it look better when compared to Apple, so they offered to intervene. Either way, GOOD.
  • Recent Achievements

    • One Month Later
      Vincian earned a badge
      One Month Later
    • First Post
      Jocimo earned a badge
      First Post
    • Week One Done
      suprememobiles48 earned a badge
      Week One Done
    • One Month Later
      Windows Guy earned a badge
      One Month Later
    • One Month Later
      Prasann earned a badge
      One Month Later
  • Popular Contributors

    1. 1
      +primortal
      509
    2. 2
      +Edouard
      172
    3. 3
      PsYcHoKiLLa
      90
    4. 4
      Steven P.
      77
    5. 5
      neufuse
      68
  • Tell a friend

    Love Neowin? Tell a friend!