Opera 10.50, the Evolution of Performance

By Vygantas Lipskas | March 2, 2010 | 28 Comments

One of our readers, Daniel Hendrycks has spent quite some time benchmarking all the weekly/nightly builds of Opera 10.50 web browser, to give you a glimpse, on how its performance has evolved.

Opera 10.50, Evolution of Performance

Here is a graph by Ichan
Opera 10.50, the Evolution of Performance

As you can see from this chart, Opera 10.50 performance was increasing with pretty much every build, making a final product significantly faster (10-20% improvement in benchmarks) when compared to first pre-alpha version.

[digg-reddit-me]


About (Author Profile)


Vygantas Lipskas is a former web designer whose projects are used by companies such as AMD, NVIDIA and departed Westood Studios. Being passionate about software, Vygantas began his journalism carrier back in 2007 when he founded FavBrowser.com. Having said that, he is also an adrenaline junkie who enjoys good books, fitness activities and Forex trading.

  • Grrblt

    Leaving nobody looking the fool. Again.

    • Bemused

      Yes, he has long been arguing that browsers get slower and slower over time…

    • nobody

      builds 3261 – 3273 were faster than final

      and final isnt polished and stabilised yet, so no reason of slowdowns.

      each and every software pays speed price for stability and polish. this sometimes is outweight with other optimisations, but mainly it isnt. but you are too young and inexperienced to look beyond your fanboi gods

      • nvm

        Sorry to have to burst your bubble, nobody, but neither Chrome nor Opera will slow down over time.

        If anyone here is a fanboy here, it’s you. Your aggressive defense of Firefox by bashing Chrome and Opera is getting old.

      • Bemused

        @ nobody – do post some benchmarks, or any other evidence showing how firefox 1 is faster than firefox 3, how opera 7 is faster than opera 9, how chrome 1 is faster than chrome 4.

        Can’t? Didn’t think so.

        • nobody

          couldnt be bothered to explain basics of software engineering to some dolts.

          and to learn you to read, cos you can see the words, but somehow cannot grasp the meaning

          • nvm

            You wouldn’t know basic software engineering if it punched you in the face.

            Has Chrome gotten slower? No, it has only gotten faster over time.

            The evidence is there. Firefox fanboys just choose to ignore it.

          • Bemused

            Sorry, you’re going to have to do better and come up with evidence, rather than spewing random insults.

          • John Cioni

            nobody: I’ll assume one thing here; you have some sort of history / active involvement with software engineering. I also happen to have 7 years of full time daily software development sometimes as a sole developer on a project. I *get* your point by stating something that seems obvious – polish & stability require refined and most of the time additional code to bring a product to a “final” phase. I think where your assumption goes south is that by adding several lines of code (or modifying several lines of code) that the extra bytes compiled down into the released product will *always* result in a slower runtime. This is simply not true. A GOOD software developer will stabilize / refine their work with proper guideline based design techniques that will more than likely NOT degrade performance. I have many products that have benchmarked better all around later in their lifespan than they did when they were raw proof-of-concept mock-ups, truthfully. Sometimes the refinements that bring stability and enhancements are actually what end up improving the performance of a product, which is what we’ve been witnessing in the browser wars for years now.

            Please stop spreading your “knowledge” of software engineering, it simply doesn’t hold water.

          • nobody

            and how many layers of middle-ware, how many different components from different vendors (internal teams) with how many internal agendas your software consisted of?

            things from small scale do not translate well into 400 man teams in few countries (with time, culture and language bariers to cross). in such cases necessary encapsulation, isolation and clearity of interfaces and intermodule api’s costs much more than few lines of code. and most errors occur between modules, not inside them (unit testing and in general good internal QA makes sure, that atom units themselves are robust). this is something much much higher-level than optimising sort routines or using efficient hashing.

            want to get rid of ‘surprises’ when one team changes something not letting you know (a sad norm), you over engineer your side, just in case. that really makes code heavier over time.

            and given the unholy god of marketing deadlines, such workarounds last far longer than all involved had hoped for.

            i know the rules of conract programming, and trusting others. but.. world isnt a perfect place.

          • nvm

            So because “nobody” was part of a team of amateur fools, all software engineers are amateur fools :D

  • http://my.opera.com/aleksander Aleksander

    Awesome job Daniel!

  • maskokot

    They did apparently good job. I hope they fix all those UI bugs and crashes fast:D

    try this if you want to crush your opera : http://niels.vg/media/opera_crash.html

    or try to disable and enable Aero while opera is running.

    • Ryan

      Yep I get the same problems with those here.

    • nvm

      Quit spamming, willya.

  • http://my.opera.com/rafaelluik Rafael

    Speed increase as getting closer to final? That no one was expecting, I didn’t think it would become so fast! – and buggy

    • nvm

      I tried it. No more bugs there than other browsers, I’m afraid.

      On the other hand, Opera might be fast at the moment, but Chrome 5 will be faster than Opera 10.5. Wait and see.

      • http://Troll.com troll

        What if it won’t ?

        • nvm

          Do you think Google is just going to sit back and let Opera be the fastest? I don’t think so. Google has a lot more money, so Chrome will be faster than Opera in the end.

          • ichann

            Do you think Google is just going to sit back and let Opera be the fastest?

            That is an assumption. How are you for certain that google will do such a thing? Given Opera’s small (seems like it is growing a bit) market share, The web(Google) can still advertise it has the fastest browser without ever so lifting a finger for a new release.

          • nvm

            It’s not just an assumption that Google is much bigger, has much more money, and can basically hire the most brilliant minds in the industry.

            Sure Google can advertise being faster. But I believe that Chrome will be faster than Opera in the end anyway without having to lie in ads.

  • Ed

    could we see this in the form of a graph? because it doesn’t really look to me from the data that “performance was increasing with pretty much every build”

    • http://www.favbrowser.com Vygantas Lipskas

      Yes, I believe.

      Daniel, can you do that? Or, if possible, send me numbers in text format, and I will do that, thanks.

  • Somebody

    Here is a great article by one of the devs at Opera. The guy admits 10.5 needs polishing but has also said that they Opera will continue their current pace of development. And seems like Hardware acceleration is coming soon.

    http://my.opera.com/kilsmo/blog/2010/03/02/opera-10-50-the-beginning

  • ichann

    After viewing that thing, I can vouch that it is missing some critical parts. I have uploaded the Excel sheet that can be tweaked to your liking (wasn’t very good with excel in the first place :P)

    http://users.tpg.com.au/ichann//Book1.xlsx

    Enjoy

  • http://www.oceanaudit.net myfreightaudit

    haven’t tried opera yet…have been loyal to chrome for 4 years now :D guess i should give it a shot this time ;D what ya think?

    • max1c

      I think that chrome sucks. Thats what I think…

«
»