A14X Bionic allegedly benchmarked days before Apple Silicon Mac event

2

Comments

  • Reply 21 of 50
    Is there a native version of GeekBench for ARM MacOS? If this is running on an ARM Mac it’s likely running through Rosetta and real benchmarks should be 25%+ higher.
    jido9secondkox2watto_cobra
  • Reply 22 of 50
    wood1208wood1208 Posts: 2,905member
    Right in your face to those(you know who you are) who said Apple should leave MAC processors to Intel/AMD. I knew Apple would not move away from Intel unless have ability to produce MAC CPU/GPU at the same level(performance,efficiency) of Intel or better. 4 more days and will find out.
    edited November 2020 watto_cobra
  • Reply 23 of 50
    entropysentropys Posts: 4,152member
    Is there a native version of GeekBench for ARM MacOS? If this is running on an ARM Mac it’s likely running through Rosetta and real benchmarks should be 25%+ higher.
    It’s an A14X so I expect it is iOS, not macos10.11
    watto_cobra
  • Reply 24 of 50
    mattinozmattinoz Posts: 2,299member
    Make the fact the iPad is still hamstrung by mobile conversions a tad embarrassing.
    PadOS needs love to make use of this speed.
    watto_cobra
  • Reply 25 of 50
    wizard69wizard69 Posts: 13,377member
    cloudguy said:
    Take a step back a bit. Qualcomm's best laptop chip (used in Windows Surface type devices and soon in Chromebooks) is the 8cx and it has benchmarks comparable to the Intel i5.

    https://www.notebookcheck.net/Qualcomm-s-Snapdragon-8cx-amasses-respectable-score-on-Geekbench-and-closes-in-on-Intel-s-Core-i5-8250U.434104.0.html

    And the 8cx came out in early-mid 2019 on the 7nm process, making it equivalent to the Qualcomm Snapdragon 855 that is in the Microsoft Surface Duo phone/tablet and last year's flagship Android phones like the Pixel 4 and OnePlus 7. Meaning that the next Qualcomm Snapdragon laptop chip - which will come out this year on a 5nm process and contain the new Cortex X1 super core jointly designed by ARM Holdings and Samsung - will match the Intel Core i7. Early benchmarks from Qualcomm and Samsung chips are in that use the new Cortex X1 design and 5nm process which state that the increase in performance over the 865 - the same basic architecture as the 8cx - is significant. 

    So the ability of smartphone SOCs to surpass AMD and Intel CPUs in clock frequencies and benchmark tests has been no big deal for quite some time. For example, Snapdragon CPUs in Android phones surpassed the 1.1 GHz dual core Intel Core i3 that is in the MacBook Air some time ago. The question is whether these ARM CPUs can have similar or better performance when running a laptop operating system - macOS as opposed to iOS - with equivalent workloads - i.e. true multitasking/multithreading and heavy I/O that Apple doesn't allow on iOS or even an iPad Pro running iPadOS. For example, running a demo full stack application that has the client app, server app and middleware in separate containers ... the sort of thing that college students in web development classes run all the time. 

    THAT is what everyone is waiting to see. The Windows on ARM devices don't give us any indication because the app support is terrible, and the first ChromeOS device on Snapdragon 8cx won't hit until 2021. So next week is when everyone is going to find out how much better Apple Silicon is going to be than i5, i7 and possibly i9 chips on professional workloads.
    This isn’t an issue if Apple properly configured the machine.   If people remember back to the early days of iPad the units barely worked on the web.   That had nothing to do with the processor but rather Apple being stingy wit RAM.   

    In the case of Apple Silicon Macs they will deliver what ever performance the benchmarks suggest.  How well they work with MacOS will be more about external factors.   Sadly we know Apple has a bad history of shipping hardware with far too little RAM and secondary storage so yeah they could screw it up.   On the otherhand they have huge incentive not to gimp the machines.  To be successful they have to demonstrate superior performance.  

    In any event I’m not worried at all about performance, the A series has already demonstrated that it can do the job.   Combine that with the fact that Apple will be producing Mac optimized processors and I’m comfortable.   At least when it comes to performance, I’m truly concerned about other issues that might cause me to reject the machines.    
    bulk001
  • Reply 26 of 50
    wizard69wizard69 Posts: 13,377member
    entropys said:
    Is there a native version of GeekBench for ARM MacOS? If this is running on an ARM Mac it’s likely running through Rosetta and real benchmarks should be 25%+ higher.
    It’s an A14X so I expect it is iOS, not macos10.11
    A14X will be perfectly fine in a Mac Book replacement and other machines.    The developer platform already runs fine and with a little fore thought into what the A14X is, it could be highly optimized.  
    watto_cobra
  • Reply 27 of 50
    Excited about Apple Desktop Silicon!
    edited November 2020 watto_cobra
  • Reply 28 of 50
    You would expect the laptop variation to have higher clock speeds due to the larger battery sizes and expectation that they would be used plugged-in more often. But Apple could also use this change as a way to reduce battery sizes across the laptop lineup and save on materials + further reduce weight. At any rate, they have more options and flexibility in those areas with the A series. 
    watto_cobra
  • Reply 29 of 50
    lkrupplkrupp Posts: 10,557member
    "The benchmark testing was reportedly performed with Geekbench 5 on an unknown device. It is unknown if this benchmark is real, and AppleInsider was unable to verify the provenance of the information with certainty.”


    I would like to congratulate AI for including the above paragraph. This clarifies the status of this rumor.

    muthuk_vanalingamwatto_cobra
  • Reply 30 of 50
    cloudguy said:
    ...
    The question is whether these ARM CPUs can have similar or better performance when running a laptop operating system - macOS as opposed to iOS - with equivalent workloads - i.e. true multitasking/multithreading and heavy I/O that Apple doesn't allow on iOS or even an iPad Pro running iPadOS. For example, running a demo full stack application that has the client app, server app and middleware in separate containers ... the sort of thing that college students in web development classes run all the time. 

    THAT is what everyone is waiting to see. The Windows on ARM devices don't give us any indication because the app support is terrible, and the first ChromeOS device on Snapdragon 8cx won't hit until 2021. So next week is when everyone is going to find out how much better Apple Silicon is going to be than i5, i7 and possibly i9 chips on professional workloads.
    LOL...   Honestly, I really don't understand this question.  I hear this echoed from time to time and it's kind of funny.  What do you think benchmarks are?  Do you think the version of something like Geekbench has different workloads from the PC version of Geekbench?  Do you think the various workloads in these benchmarks don't represent common or even professional workloads of desktop apps?  Perhaps you should take a look the different workloads in Geekbench 5, then tell me how they don't apply to desktop apps, etc.   https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf

    It's so funny, some people seem to think that a processor architecture that was initially used for a phone couldn't possibly be used in a desktop.  One would think that maybe taking a look at what people are saying about the older A12z in the developer kit would remove such doubts.  I suppose there is a doubting Thomas in every crowd.  I suspect you'll get your answer next week with a few performance demos as well.
    thtwatto_cobra
  • Reply 31 of 50

    I don’t put much weight on rumors but those numbers looks good. I’ll wait for the real thing then maybe replace my 2012 15” MBP. 
    Yeah, it's best to wait until next week and then we'll know for sure.  Still, these numbers are absolutely inline with expectations for the A14x.

    mjtomlin said:
    tipoo said:
    I'm giddy, I can't wait for the event and results. 

     Gurman says the first Mac parts will be 12C, 8 big/4 little, so I think this 8C part won't be it, unless they are only counting big cores. Or else A14X for the 13" Air and/or 12", this 12C Firestorm/Icestorm part for Pro? A14T will be that, perhaps. 

    I still think the A14X is for the iPad Pro and the A14T for a new gaming-centric Apple TV. (The A14T was mentioned alongside a new custom GPU.)

    Mac SoCs will have different branding to differentiate them from iOS devices, even though they'll most certainly be "based" off the A-series, i.e. use many of the same IP blocks. I think Apple's goal is to get the Mac line completely above (or match) the performance of the iPad. Seriously, it is embarrassing when a phone you put in your pocket is almost twice as powerful as your laptop! So Apple will want to draw a performance line across all their devices. Where performance of the top iPhone mixes with bottom iPad, and top iPad mixes with bottom Macs.
    If rumors are to be believed, the bigger A14T to be used in iMacs won't be out until the first half of 2021.  The A14x will be plenty powerful enough to get the party started with laptops.  iMacs will come in spring / early summer.  As for the Mac Pro... who knows.  That may be another year or so away.
    watto_cobra
  • Reply 32 of 50
    KITAKITA Posts: 392member
    techconc said:

    Do you think the version of something like Geekbench has different workloads from the PC version of Geekbench?  Do you think the various workloads in these benchmarks don't represent common or even professional workloads of desktop apps?  Perhaps you should take a look the different workloads in Geekbench 5, then tell me how they don't apply to desktop apps, etc.   https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf
    Just a note from Anandtech in their recent Zen 3 deep dive: 

    Most of the people in our industry have a love/hate relationship when it comes to synthetic tests. On the one hand, they’re often good for quick summaries of performance and are easy to use, but most of the time the tests aren’t related to any real software. Synthetic tests are often very good at burrowing down to a specific set of instructions and maximizing the performance out of those. Due to requests from a number of our readers, we have the following synthetic tests.

    ...

    GeekBench 5:

    As a common tool for cross-platform testing between mobile, PC, and Mac, GeekBench is an ultimate exercise in synthetic testing across a range of algorithms looking for peak throughput. Tests include encryption, compression, fast Fourier transform, memory operations, n-body physics, matrix operations, histogram manipulation, and HTML parsing.

    I’m including this test due to popular demand, although the results do come across as overly synthetic, and a lot of users often put a lot of weight behind the test due to the fact that it is compiled across different platforms (although with different compilers).

    muthuk_vanalingam
  • Reply 33 of 50
    cpsrocpsro Posts: 3,192member
    I had hoped Apple Silicon at 5nm with twice the general purpose registers would do better speed-wise than AMD Zen3 at 7nm, but it would appear to be about the same on a core-to-core basis. Power-wise, Apple almost certainly does better than AMD (and will absolutely crush Intel).
    edited November 2020 bulk001
  • Reply 34 of 50
    KITA said:
    techconc said:

    Do you think the version of something like Geekbench has different workloads from the PC version of Geekbench?  Do you think the various workloads in these benchmarks don't represent common or even professional workloads of desktop apps?  Perhaps you should take a look the different workloads in Geekbench 5, then tell me how they don't apply to desktop apps, etc.   https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf
    Just a note from Anandtech in their recent Zen 3 deep dive: 

    Most of the people in our industry have a love/hate relationship when it comes to synthetic tests. On the one hand, they’re often good for quick summaries of performance and are easy to use, but most of the time the tests aren’t related to any real software. Synthetic tests are often very good at burrowing down to a specific set of instructions and maximizing the performance out of those. Due to requests from a number of our readers, we have the following synthetic tests.

    ...

    GeekBench 5:

    As a common tool for cross-platform testing between mobile, PC, and Mac, GeekBench is an ultimate exercise in synthetic testing across a range of algorithms looking for peak throughput. Tests include encryption, compression, fast Fourier transform, memory operations, n-body physics, matrix operations, histogram manipulation, and HTML parsing.

    I’m including this test due to popular demand, although the results do come across as overly synthetic, and a lot of users often put a lot of weight behind the test due to the fact that it is compiled across different platforms (although with different compilers).

    Kita, I'm not sure what the point of your post is. I know that Geekbench is a synthetic benchmark. It's also based on practical workloads. If you read the document I linked to, that would be clear. What would also be clear is that the comments on Anandtech are not correct regarding compilers. They use the same Clang / LLVM toolchain across platforms. Again, this is outlined in the document I linked to.
    thtwatto_cobra
  • Reply 35 of 50
    wizard69 said:
    entropys said:
    Is there a native version of GeekBench for ARM MacOS? If this is running on an ARM Mac it’s likely running through Rosetta and real benchmarks should be 25%+ higher.
    It’s an A14X so I expect it is iOS, not macos10.11
    A14X will be perfectly fine in a Mac Book replacement and other machines.    The developer platform already runs fine and with a little fore thought into what the A14X is, it could be highly optimized.  

    Yep!!!
    watto_cobra
  • Reply 36 of 50
    Dang. 

    If the A series is trouncing Intel i9 CPUs already, desktop class M series or whatever is going to be crazy. 

    This is already crazy. A mobile chip smashing Intels premium desktop chips...What a day we live in. 

    Now think desktop class chips... the future of Apple Silicon is bright. 

    Steve Jobs would be proud. 

    Mr. Cook has Proven to be a more than excellent choice to helm the ship that is Apple. 

    I remember the day when people thought Apple was over after Jobs. Cook has taken the foundation and ethos and brought it to new heights. 

    Wow. 
    watto_cobra
  • Reply 37 of 50
    this result is something that I would never expected early this year when leakers said there's gonna be new Apple Silicon-based Macs, I thought it would be MacBook Air or 13 MBP the most, but looks like they could easily make top-tier 16-inch MBP with this horsepower. Absolutely amazing!
    watto_cobra
  • Reply 38 of 50
    The performance, if true, is rather an evolution and not a revolution. Which is fine.

    However, the BIG change to me would be this chip with no active cooling required. And/or an Apple GPU that can match mid to high-end discreet graphics from the likes of AMD or Nvidia.
    edited November 2020
  • Reply 39 of 50
    red oakred oak Posts: 1,087member
    This puts it at the very top of Mac Pro 16” 
    performance 

    Awesome 
    watto_cobra
  • Reply 40 of 50
    d_2d_2 Posts: 117member
    The semi obvious thing that isn’t being said here is that Apple has *complete control* of the processor design and OS for the new Macs, like they have had for iOS devices... thus, besides the usual compute logic within the CPU, there are all of the other optimizations that they are building into the silicon for Big Sur to take advantage of - many of which are known / have been explained... but, what else might Apple be planning?
    watto_cobra
Sign In or Register to comment.