Opera Says Its Browser Runs 2X Faster on M1

Posted on April 1, 2021 by Paul Thurrott in Mac and macOS, Web browsers with 12 Comments

Opera today announced the availability of the first M1-native version of its flagship browser, and it’s claiming that it’s two times faster than the previous version.

“When you buy a shiny new Mac, one of your key priorities is speed: you want it to load and run apps as smoothly as it can, and you want to be impressed by how seamless the experience is,” Opera’s Julia Szyndzielorz writes. “While marveling at your crisper than ever display, you also want to browse the web with freshness and speed inspired by your brand spanking new Mac. Apple wants you to be happy, and so does Opera, which is why we’ve made your browsing on Macs with M1 even faster.”

It’s not clear where Opera got the 2X number. Other companies, like Adobe, only claim significant gains of up to 1.5X in certain operations while noting that everything “seems” faster. But that’s not the only change in the new browser: Opera says that it also added customizable keyboard shortcuts for Flow, including those for Opera’s built-in Crypto Wallet.

You can learn more about the changes in Opera 75 in the changelog.

Tagged with ,

Join the discussion!

BECOME A THURROTT MEMBER:

Don't have a login but want to join the conversation? Become a Thurrott Premium or Basic User to participate

Register
Comments (14)

14 responses to “Opera Says Its Browser Runs 2X Faster on M1”

  1. Avatar

    F4IL

    > It’s not clear where Opera got the 2X number.


    The previous version was probably running under Rosetta.

  2. Avatar

    richardbottiglieri

    My experience with Firefox on my M1 MacBook Pro versus my Intel i9 Mac has been that the M1 version is faster. It feels faster, and benchmarks are roughly 50% faster. This is the M1 MacBook Pro versus a 3.6 GHz Intel i9 iMac with 8 cores.


    I use Microsoft Edge more on both computers, and that is also faster on the M1 than it is on the i9 iMac.

  3. Avatar

    rosyna

    Apple also implements ARMv8.3, which includes an instruction to significantly optimize an exceedingly common JavaScript operation that’s otherwise very slow. If a JavaScript engine implements this instruction, the gain can be huge.

  4. Avatar

    IanYates82

    Do they mean 2x faster than it being emulated under Rosetta, or that the M1 version is 2x faster on the M1 than an x86 version on an equivalent Intel chip?

  5. Avatar

    madthinus

    Someone drank too much cool-aid. Think the bulk of these speed ups have nothing to do with the cpu speed, but rather code optimisation to better suit RISC compilers.

  6. Avatar

    nbplopes

    Technical people in the active know and understand the speed gains on the M[ compared to Intel devices in the same range.


    Its not rocket science. The first impact can experienced immediately upon using it as everyone can testify, including of course Adobe.


    Some cases is 50% faster which is huge considering that desktop and laptop CPU mostly flatted out in the last 5 years, other times it’s 2 times or even more performance gains for the same class, which is outstanding.


    As far as I can tell, the consensus amongst developers is that this era starts with an entry level device that in performance mostly equivalent to a high end device in the previous era. This as far as most devs activities go.


    The current core drawbacks is that it comes with some compatibility issues that might be show stoppers for now. Android simulators for testing .... Electron apps work extreme fast, no need for snapshots. Give it another year or so for these to be totally ironed out


    • Avatar

      bkkcanuck

      In reply to nbplopes:

      The M1 should be measured in the class that the processor was developed for... which is a fanless entry-level machine (ipad of laptops; or the MacBook 12" in MacBook Air form-factor) [the M1 Mac Mini exists to temporarily replace the Mac Mini 'developer edition', and the Macbook Pro is a placeholder until the chip for the MacBook line (maybe two versions) exist]. In the end, it will be the worst performing M* Mac ever.  


      I think there is potential that the next set of Macs will be based on the M2 core architecture (which I suspect will be based on the ARM v9 release). The ARM v9 architecture will bring Realms and tagged memory (both great improvements in security architecture) as well as Enhanced Vector Processing (introduced as necessary improvements for the Fujitsu Fugaku Supercomputer - i.e. further improvements in Matrix and Vector). ARM v9 includes full backwards compatibility with ARM v8 - so it should be a slide-in replacement.


      The M1 Mac slaughters the Intel CPUs when it comes to efficiency - and it is a rather small CPU (119 sq mm) [the GPU for nVidia is 3080 is around 620 mm sq. - so there should be room for improvement]. Yes we can only speculate what Apple will release, but given all the pieces that we can see... Apple has a lot of headroom that Apple has available to improve. There is a lot of thermal headroom (M1 works quite well with no cooling), the single core performance is outstanding, there is room for Apple to quadruple the number of performance cores as well as graphics cores [maybe not quadruple both in the next version of the Macbook though]. Apple will likely start using 4nm fabs this year, and some limited use of 3nm (so further performance or efficiency improvements from fabrication improvements). All this bodes well for signnificant improvements continuing for the next several years.


      A good 60%+ of Applications have made the transition to M1 already, and as the full MacBook and iMac lines being released sometime this year (with Maybe the Mac Mini Pro / Mac Pro next) -- the pressure to transition the rest of actively supported software will not be able to be ignored. (I suspect Rosetta has a maximum of 3 or 4 versions of macOS before it is retired again).


      It will be very difficult for Intel to catch up to what Apple is doing now and probably impossible compete with technically speaking with what I believe is likely coming fairly quickly.

    • Avatar

      luthair

      In reply to nbplopes:


      Note - they are claiming it is 2x faster than the previous version - i.e. the x86 translated version which shouldn't be particularly surprising. This has nothing to do with intel vs m1 performance.

      • Avatar

        bkkcanuck

        In reply to luthair:

        No, but the M1 Macbook Air has a CPU benchmark performance increase of last Intel Macbook Air 2020 [Intel Core i7-1060NG7] (same thermal body but with a cooling solution)... of 138% (based on multi-core performance -- Geekbench 5 Benchmark)... (Sidenote: the M1 has over a 400% improvement in Metal Compute over the same previous Macbook Air 2020)... Of course overall performance improvement is dependent on all hardware (SSD, memory, etc.) ...

  7. Avatar

    ivarh

    Since the limiting factor from a web browser is going to be the speed of your internet connection I doubt you will feel much difference between a browser running under rosetta (2) and a native compiled browser. What you will notice is the power usage. When I use my M1 MacBook pro under normal use (RDP/Zoom/Webex) I see the CPU drawing around 2 watts according to the activity monitor on my mac. This machine does not get warm. This is the first laptop I have had that is lappable without becoming uncomfortable warm regardless of what you do. During the last month, I have been using my M1 MacBook pro working from home and with a 100% battery at the start of the day, it is around 60% after 8-9 hours with the screen on 100% of the time. And it never feels warm. I have to admit that apart from my work provided dell laptops I have not used a windows based laptop for the last 10 years, instead running my work windows image in a VM but only after I swapped to a M1 based mac have I had a machine that has been comfortable using in my lap for long periods of time. This includes the dell laptops that work have provided me with.

Leave a Reply