Win; 
Mac; 
Linux 
Opera is obviously not content with sitting on the sidelines when it comes to being the fastest browser out there. Back in February Mozilla decided that it was time to optimize it’s JavaScript performance, and used a technique called Profile-Guided Optimization (PGO) that gave it a nice boost. Opera decided to jump on the PGO bandwagon as well with this week’s release of Opera 9.5, and here is their explanation of how it works:
With this technique, Opera is compiled twice. The first compilation creates a specially instrumented build that gathers statistics about which code is called and which variables are used most often. The instrumented build loads sites automatically using “URL player”, a testing feature available in every Opera build (more information about that in a later blog post). This information is dumped into files and used to produce the final build that’s optimized based on the gathered statistics.
Did it help? Unfortunately there are some bugs in Opera 9.5 that prevent it from finishing the Sunspider test, but in the areas it does complete the result is normally better than Firefox 3 Beta 5. It does, however, consistently beat Firefox 3 in the MooTools SlickSpeed test in addition to this JavaScript benchmarking tool.
Congrats Opera! It looks as though you might be recapturing the JavaScript performance crown, but we’ll have to wait and see what the final version of each browser is like before we make any conclusions along those lines.
Be sure to checkout our more complete article where we compare the performance of more browsers.
Enjoyed the post? Subscribe to our feed to get a daily dose of CyberNet!
Tags: Freeware, Pre-Release, Software, Beta, Browser, Firefox 3, Opera, Performance
Related Posts:
- Opera Takes Home The Gold In The Browser Speed War
- Opera Dragonfly Developer Tools
- Making Google’s Services Work with Opera
- Built-In Opera Features That Firefox Needs Extensions For
- CyberNotes: Crushing Those JavaScript Pop-ups In Firefox
So, let me see if I’m clear here. Opera broke their browser by optimizing it and now it’s broken on some tests but on some tests it’s faster than Firefox 3 which isn’t broken on any tests and that translates to “Opera beats Firefox 3″?
Opera has been the fastest browser for a while - but they really need to work on compatibility. I try new versions of Opera when they come out, but usually ditch them within an hour because some web app won’t play nice with it.
I successfully ran Sunspider benchmark on both Opera 9.5 build 9981 and FF 3.0 beta 5. Opera scored 5853.6ms +/- 1.0%, while FF scored 5280.6ms +/- 6.0%. If you take the margin of error into consideration applying -1% to Opera and +6% to FF we get 5795.064 for Opera and 5597.436 for FF. No matter how you stretch the benchmark, FF takes the crown, but not by much anymore. You can definitely see the improvements in Opera’s JavaScript engine thanks to PGO.
BTW, I ran this on Penium M 1.5GHz with 1GB or RAM if it matters to anyone. Also, I upgraded my Opera installation to latest build from the first beta 2 build.
StationStops, your claim would seem to be contradicted by all the available evidence.
See Marko’s comment above. and any contemporary published benchmark.
Those are good news, every little improvement to Opera’s speed is welcome, but I hope they fix those compatibility issues, I don’t want even less site fully supported with Opera
I guess the optimizations were done to help Dragonfly’s performance as well.
Speed is not the only consideration when choosing a browser. Sure it’s nice if a browser is fast, but you also have to look at how you work and what fits best with the way you work. For me FF works best for me and the way I work. However I have found one instance where Opera works better for me and that is a portable app to carry around on my flash drive; Opera@USB http://www.opera-usb.com/operausben.htm works great for this. In that instance FF lets me down due to too much writing to the flash drive which really makes it to slow to be usable. The fact that this new version breaks so much will probably keep me away from it for now. Thanks for the update on the browser speed wars though; it always interesting to see where the browsers are headed.
It is still a Beta version of the browser, and that has to be taken into consideration. Sure it didn’t run all of the tests, but if it did then it’s likely that it would have beaten Firefox on those, too.
Yeah, the compatibility thing is a big downfall. But what sucks for them is that they aren’t the ones completely at fault. It’s also developers who don’t care enough to make their site play nice with Opera.
Thanks for the info! I wonder why I couldn’t get it to successfully run on three different computers?
StationStops is actually correct in several ways if you only look at browsers that have been released.
I didn’t even think about that, but I bet you’re right. Dragonfly uses all JavaScript so it will definitely benefit from this.
That’s a really good point, and excessive writing to the Flash drive can actually shorten the life of the drive.