anonconformist

About

Username
anonconformist
Joined
Visits
111
Last Active
Roles
member
Points
585
Badges
0
Posts
202

Reactions

214Like2Dislike75Informative

  • How Apple A-series chips stack up against Intel Macs

    anonconformistanonconformist
    chasm said:
    What Metriacanthosaurus said. While some new benchmarks on the 12Z chip running native apps suggests that even that chip has an impressive amount of barely-tapped potential, the real point here is that Apple has not yet released any member of the "family of SoCs" that are specifically designed for the Mac. Nor has it yet released any device that has all of its systems optimized for the (presumably redesigned) future Macs.

    Recent stories here and elsewhere are kind of focused on the chips themselves because that's where the (tiny amount of available) information is right now, but there's really four keys to future Macs: first, the specially-designed Mac-centric SoC; second, the optimized motherboard and other mechanical components (we've been enjoying some of this part for years now, which is why we're so far ahead of PC makers on things like TB3/USB-C/USB4); third, software compiled specifically to take advantage of this custom hardware (in particular, I foresee much greater expansion on multi-tasking for all but the most basic software); and four, new chassis designs again fully in harmony with the chip, graphics, thermals, and support systems.

    Intel has been great for Apple, and often very accommodating to Apple's specific desires (especially given the size of the Mac market, which is still small). But Apple Silicon (please note: not silicone!) is made not only with an awareness of hardware requirements but also the specific ways Mac users work with software, neither of which Intel can really design for. Because of the potential this opens up, I think Apple will continue to work on bringing more components in-house (like wireless/5G) and forming partnerships with third-parties where they can get custom-designed parts for its particular and in some cases unique needs.

    Kudos to Mike P for the rundown, though this is really only the beginning of a tale mostly yet to be revealed. Specs and comparisons are fun, but at the end of the day the truly remarkable thing about all this is that Apple has been planning a brilliant revival of the Mac -- by far their least-popular "computer" product, let's not forget -- to keep it relevant and even exciting in the current age of mobile and wearable domination. I cannot wait to see what Apple Silicon running new Apple hardware can do, how much room for innovation they've opened up, and specifically (though I'm not the target for it) what high-end Apple Macs are going to be able to do. I'm with Jean-Louis Gassée in thinking that some major surprises are still ahead of us, including some that will benefit the entire industry, directly or indirectly.
    I’ll address one point: you’re clearly not a developer, or at least not one with enough experience.  If you make a faster CPU it doesn’t change a thing about what you can do regarding multitasking, regardless of single core or multicore performance, and adding more cores does nothing to change the nature of how you solve problems effectively by adding more threads in a process or more things in the system as a whole.  In fact, adding more cores and more multitasking/threads slows down throughput for a process and often a system, when those related tasks aren’t embarrassingly parallel that don’t have memory conflicts.  Look up Amdahl’s Law, and what’s important to understand is even the best system is constrained by main memory latency and speed, with the CPU waiting for data from main memory about 30-40% of the time, despite all the huge cache hierarchy: it’s made far worse if you quickly swap unique code in and out of the CPU for as many threads as possible, as quickly as possible.

    There are no applications and domains that exist that don’t already work well on a high-end MacPro and not-quite-as-fast on an i9, but adding more cores doesn’t automatically enable the developer to accomplish anything, and it isn’t any different in how it is designed by the developer.
    fastasleepFileMakerFeller

  • TSMC 3nm 'risk production' in 2021 paves the way to 2022 mass production

    anonconformistanonconformist
    bageljoey said:
    rob53 said:
    There has to be a limit on how small you can go before the size of the components won't work anymore. Once this limit is reached, how will processor fabrication change to improve speed and reduced power usage? 
    I always recall writing a paper for a chemistry class about computer chips (back in the late 80s). The books all stated as a fact that the laws of physics limited processors to 100nm but that 200nm was probably the practical limit. 

    I believed that even when they got to 100nm and I take everything as magic since then...
    i no longer believe there is a limit!
    There are theoretical limits, and those can more readily be computed.

    We are certainly approaching silicon limits.

    Even with other than silicon being used, theoretical limits say we can get this small, but how quickly and at what defect rate can that be done?

    There is also the design time to most effectively use all the (theoretically available) transistors, as it's not quite as simple as copy/paste another CPU core with identical connecting logic in between.

    Finally, there is the cost of developing and building new manufacturing tech and facilities: the smaller the process node, the more expensive it gets, by quite a bit.


    viclauyycrmusikantowwatto_cobra

  • Should you wait for Apple Silicon to upgrade to a new Mac?

    anonconformistanonconformist
    johnbear said:
    History will repeat, Apple failed at CPUs before. 
    Grab one with intel while you can before the Mac becomes an iOS device.
    If you were honest, Apple hasn’t failed at CPUs before, their suppliers have.

    Who designed and manufactured all the Motorola 68k and following generations? Motorola: Apple had no control of design/manufacture of them.

    Who designed and manufactured the PPC processors? A combination of Motorola and IBM.  Apple didn’t design or manufacture them.

    Who designs and manufactures every Intel processor? Intel: pay attention, there’s a clear pattern here!

    Since the Apple A4, who designed the CPUs (part of the SOC) in every non-Mac device since that time? Apple!  What’s their performance been? Perpetually increasing in both power efficiency and CPU throughput.  Apple has been able to improve them in every single generation at a schedule of their choosing, and they’re currently being manufactured on a process better than Intel’s and Intel is at least a couple process nodes away from being a match on that alone.

    Now, will Apple eventually hit a limit?  Will Intel/AMD hit a limit? There are data points from either side that says it could go either way and performance improvements stall.  There’s nothing at all to say the most efficient ISA, given identical process nodes, are either of those at all: I’d wager there’s an architecture/ISA not yet on the market that can be pushed to more processing power at less power usage.  If a clear winner for optimization becomes apparent, Apple at least has a lot less inertia in changing over to a completely new ISA because Apple (for better and worse, as Apple is a consumer electronics company) doesn’t come close to having the concern for backwards-compatibility that constrains Windows from faster change, but makes it well-suited to be the OS with the majority of the market, where Enterprise is a huge money-maker factor.
    mtlion2020cornchipfastasleepwatto_cobra

  • Apple's macOS 11 Big Sur marks the end of OS X, not the Mac

    anonconformistanonconformist
    Fatman said:
    The current Mac App Store is limping along - this change will breathe new life to Mac apps, the lifeblood of any OS. If MS Office (needed for legitimacy), and ‘big screen’ oriented Pro Creative Design, Audio and Video apps smoothly make the transition, then the platform will continue to live on and even grow beyond its stagnant 10% market share. It is critical for Apple to support developers in porting to ARM, and it appears they have a solid plan that is already quite far ahead. Developers targeting iOS can now leverage their work for the Mac and expand their audience (and revenue) for a limited amount of effort. What will really be attractive - when Intel chips are shed - is a desktop/laptop platform that is more cost effective (think more $999 Macs - not $3000 Macs) and that is incredibly fast (potentially 2 or 3 times faster than Wintel Systems). Two advantages Mac needs.
    I’d be curious what brings you to the conclusions of 2-3 times faster and $999 Macs with going to Apple Silicon: while Apple can almost certainly save quite a bit of money per chip (especially via making it SoC except for RAM, SSD and often discrete GPU, perhaps a few other things) just throwing in a fast  CPU alone does not guarantee a fast system: you still need high-speed SSD and RAM subsystems to keep the hungry beast fed.  Those things cost money, and a CPU 2-3 times faster needs notably faster and more expensive memory to achieve that.  The price of memory that fast aren’t likely to only be that much more expensive, but nonlinearly more expensive.

    I’ve watched the CPU benchmarks of the A series CPUs over the past few years getting into Intel x64 levels of performance, in battery-powered passively-cooled systems, and wondering what they could do with more capacity to dissipate waste heat and not stuck on tiny batteries, but also keeping in mind Apple has tightly designed their systems for maximum bang for the buck with balanced systems.  Compared to desktop/laptop machines, Apple’s iPhones and iPads have always (until most recent iPad Pro with 6 GB RAM) had notably less RAM than their Macs: the more RAM you have, the more quickly the battery is drained even in standby, let alone in active use.  This RAM is not going to be the cheapest RAM because of the trade for fast-but-energy-cheap performance tradeoffs.  Put enough (16 GB these days should be the base level for a Mac) and the expense grows quickly.

    the majority of the time a CPU is executing code, 30-40% of the time every CPU core is idle, waiting for data to be retrieved from main memory.  Throwing more cores on a chip to a point ensures there are always CPU computations going on in at least one core more often than not, but not even that is guaranteed.  Adding more cores beyond a certain point absolutely guarantees the main memory bandwidth will be incapable of keeping any of them truly busy doing computations: wasted silicon, you might as well not have added those cores, because beyond a certain point, cache coherency and other factors guarantees more power wasted with lower total throughout per core and for the whole system.

    any system 2-3 times faster than Intel’s for throughout will cost a lot more than the system you compare it to in this situation.  You can’t make it all that much faster while making it cheaper, because the CPU monster Must Be Fed, and it is expensive.
    [Deleted User]bakedbananaswatto_cobra

  • Why the Mac's migration to Apple Silicon is bigger than ARM

    anonconformistanonconformist
    commentzilla said:
    rain22 said:
    rain22 said:
    “ but it suggests that new Apple Silicon Macs will not be struggling to keep up with the graphics on Intel Macs.”

    That would be nice - but seems extremely dependent on programs being optimized. The anemic library of titles will probably shrink even further - at least until there is market saturation. 

    Mac users will be stuck using dumbed down iOS software for a long time I feel. 
    After all - This is the motivation isn’t it? Eventually have just 1 OS that can be modded to facilitate the device. 
    iOS and macOS share the same core and were designed to be processor independent. This is why existing apps can be modified and recompiled for ARM in days. I suspect the performance will be far better than you think. That Tome Raider demo spoke volumes. A brand new 2020 13" 10th Gen i7 Quad-Core can't even do that with native code.

    I hope you are right. We got burned on a bunch of G5’s during the last switch. Rosetta didn’t work half the time, crashes, software wasn’t supported, and Apple dropped support on its own suit almost overnight. Peripherals became junk as no drivers were updated and Apple put the whole onus on developers and manufacturers. We ended up tossing the G5’s for a big loss and getting the new Mac Pro’s. At least then we could upgrade our own video cards.
    A lot of difference between then and now. Apple has been running iOS (macOS) on ARM for a decade now and operating has been designed to create a layer of abstraction from the hardware. This is why many apps will only need to be recompiled or can run in translation. That Tome Raider demo was a x86 mac app running without modification in translation via Rosetta 2. The old Rosetta couldn't do anything like that since apps were much closer to the hardware in those days and Metal didn't exist yet.

    As for INTEL updates, I assume any app written natively for ARM can be recompiled for INTEL which means developers will easily be able to support both platforms. That's the beauty of the abstraction layer. The only question then becomes how long will Apple support new versions of macOS on INTEL since that occurs at the hardware level. I suspect it depends on the install base and their traditional obsolete/vintage status for hardware; 5 years of full support and 2 additional years minimum for security updates.

    5-7 years is just about what I expect to get out of a device. Come this Fall my 7+ year old 2013 15" MBP will no longer be supported by the current operating system, which means it's down to security update status. I got my money's worth and the resale value of these machines is likely to remain high since Bootcamp is gone forever.
    Any OS ever meant for mainstream use over more than the last 20 years has had sufficient hardware/software abstractions built-in that it’s trivial to move the OS over to a new CPU and hardware configuration: anything else is a hobby or specialized embedded system.

    MacOS from before it was going to be used for the basis of MacOS from 2000 on, in all its variations, has been abstracted from hardware dependencies: it’s a variation of Unix, which has a history of that from early on.

    Windows NT and later (first released 1993, the core of Windows 10 today with evolution as any OS has over time) was also hardware-abstracted by design from the start, with Windows 95, Windows 98 and 98se and Windows ME being short-term backwards-compatible OSes that were very much x86-dependent by design: lots of assembly language in them to make them fit in hardware constraints, and also old compatibility from Win16. These OSes were an intentional stopgap for most home users to be able to transition without losing too much backwards compatibility as a main focus (also applied in business as well).
    patchythepiratewatto_cobra