What’s Left in the Apple Silicon Transition
512pixels.net
512pixels.net
- A flagship product that demonstrates just how powerful Apple Silicon can be. This in turn could enhance the Apple / Mac brand for a certain group of users.
- An aspirational product for users who think that they might need more power in the future.
- A means of keeping companies that need a small number of powerful machines in the Apple camp.
- A proving ground for technology that ultimately makes its way into volume products.
All of these factors could make the product justify its development costs even if actual sales are small. I doubt that the highest end models of some premium car brands recoup their development costs but their value to the brand offsets those costs.
Halo cars like Corvette Z06 are exactly the right analogy. They sell at most 8000/year; there is no way it’s profitable as a line item. But it gets made for all of those indirect reasons.
The danger there is what if it isn't? Apple's riding high right now but what if AMD or Qualcomm or whoever takes the next crown? This is one of the things that almost killed the company back in the G4 days. Jobs would be up there on stage essentially lying to our faces about the "Personal Supercomputer" when any of us could look at the literally-twice-as-fast Coppermine boxes on our desks to see the truth.
Computers and technology are ephemeral and quirky. Apple's brand isn't and has never been[1] about selling "chips" or "computers". They sell "modes of interaction".
[1] The last pure-tech play from Apple was the Apple II+.
Edit - this is a comment about branding ie how Apple presents itself not actual performance.
But the actual machines have always been merely "good enough". The 68k macs lagged PCs significantly. PPC had a brief moment where it was faster than 486/P5 devices but then got demolished by P6+ cores, to the extent that Apple had to switch to Intel themselves. Then for almost two decades they shipped what amounted to the same silicon as high end windows laptops. MacBooks were "high performance", but not in a way that wasn't accessible to HP or Lenovo.
But all that relies on it not being a lie. If you lock your platform marketing not into merely "Insanely Great" but into the idea of Apple Silicon being best, then you need to make damn sure it actually is. And right now it is! But it needs to stay that way if you're going to bet the platform on it.
Exactly the point I made.
I've got a Mac Mini M1 as soon as it came out.
I recently got a Ryzen PC, 6 core 5600G w/ 32GB RAM and 1TB SSD. Cost approx £500. Upgrading the M1 spec from 256GB SSD to 1TB (£400) costs basically the same as the entire PC I bought. I'm impressed with the 5600G, it is super cool and quiet compared to what I was expecting and the 32GB of RAM makes much more of a difference to performance than the extra CPU perf of the M1.
It's really a bit ridiculous that Apple is still selling products with a base config of 8/256. I'm pretty sure that hasn't increased since I got the retina macbook pro in 2012 unless I'm mistaken?
They were still using 128GB SSDs in the entry-level 13" MBP from 2015 (replaced by the touchbar MBP in late 2016, but not actually discontinued until sometime in 2017).
On a MacBook PRO no less.
I made no comment about actual performance.
Apple has publicly stated during their keynotes (albeit with slides that present minimal information) the efficiency their CPU holds over the aforementioned companies. And that isn't even going into the mobile space where Apple still holds the crown when it comes to performance on synthetic workloads.
Revisiting the keynote where the M1 was introduced [1] at 8:48 Apple states that the M1 uses the world's fastest CPU, which ties back into your argument, but then at 9:51 talks about performance per watt.
Apple does it again when the M2 was announced [2] at 58:21.
This isn't to say your argument isn't a valid one, and I hope I didn't give the impression that it is wrong; the optics has, at least in my perspective, shifted towards preferring power efficiency over raw performance output. What with tech news of x86 CPUs and GPUs from Nvidia and AMD consuming more and more power the M1 and M2 sip power by comparison.
No one doubts the performance of Apple devices, it's mainly efficiency by which it does so that has everyone paying attention to it.
Macs have never been about highest possible performance at any price point. But there is a niche where people want a very fast computer that is quit and not too ugly (and who can find their apps on macOS).
It's a valid approach in theory but it seems pretty clear that the market hasn't embraced it. The most recent Mac Pro recognized that people weren't embracing it, but really, are there many (or any?) 3P cards that can be plugged into a Mac Pro?
Putting two GPUS in a tight spot and the limitations of the time basically doomed it to failure but now we have much better data rates and better iGPUs.
You can build/get quiet PCs too. Theres a whole PC accessory company called be quiet! even. And there's some very nice IMO cases that look better than the Mac Studio or Pro...
My comparison is up to date as the top end chips from Intel and AMD and the top-end Nvidia GPUs handily outpeform the top-end silicon in the Mac Studio's highest end configuration. And for a lower cost. The Studio is a solid machine, but it doesn't have the same 'this is about as good as you can get without a server farm' performance of the top-end Mac Pro when it launched.
The minute however that this starts to sag they'll start cost-cutting and they'll start with unprofitable vanity products like Mac's.
NOBODY has transitioned to an other architecture like Apple has, nor has the industry been kicked in the shorts as many times as by apple. (wholesale inclusion of GPU in the OS, migration from Motorola, to Power PC, to Intel, to Arm to X Efficienty + Y Performance Cores...and nobody wrings as much out of the ARM architecture as Apple.
It's 'the same OS' from the watch, TV, Airpods to the Mac Pro.
There is no cost cutting the top end because it's ALL the same Platform. The Top end is the same as everything else with an additional parlor trick.
For the phrase "Apple is Dooomed!", including similar results, I get only 4 pages of google search results.
For "Apple is Doomed!", including similar results, I get 31 pages.
What chance can anyone else have if the Mac Pro is doomed if iPhone sales dip?
Although revenue for Mac is still increasing (most likely because they're making even more profit by using their own CPU) the sales are already softening.
History would have taken another path without that reverse acquisition and Microsoft's money injection.
More of an issue now and in the coming years is bandwidth to do work on low volume products. They may be happy to spend the money but diverting key staff who could be working on a volume product, say the next iPhone SoC, that's a tougher call.
Increasingly though it feels like those of us who prefer a laptop (or desktop machine!) over a phone are like HAMs ... or perhaps guys that like manual transmissions.
It seems more likely that it will be the Macs that go down, and not the iPhones a decade or so from now.
This is because platforms have three advantages. The first is a user’s sunk investment. I own hundreds of pounds worth of iOS software, and even the free software I use took an investment of time and effort to learn. Plus I now have considerable assets in those file formats and storage systems. Switching to a new platform would be expensive in money and time. The second advantage is network effects, my whole family is on the same platform, so is my company and many of my friends using the same services. The third advantage is the ecosystem. There is a plethora of software and services available for iOS that a new entrant won’t have, many offered by third parties. In competing with iOS and Android you’re not just competing with Apple and Google, but also the entire ecosystems of companies that complement them. These three advantages also apply to Android and its users.
This is why even Microsoft could not break into the modern mobile smartphone platforms market, they were too late. Android and iOS had critical mass in all three areas. Google was incredibly lucky they managed to get Android up to being a viable platform before Apple had the entire mobile platform market sewn up. Fortunately the Android team at the time knew the new game they were playing when they saw the iPhone announcement, and realised they were in a race against time for survival.
This is also why the desktop is dominated today by the same platforms it was in the 90s. Mac and Windows. Arguably we could extend that to the 80s with Mac and DOS. That’s around 3 decades of stability. The mobile market in its current form was established around 12 years ago. I would not be at all surprised if the mobile market is still dominated by Android and iOS in another 20 years. There are incredibly powerful forces driving this.
My surprise is that Android is the "other one" and not Windows Phone.
I'm not. Microsoft, with WP, copied the weak things from iOS (like being locked down, but also with very poor walled garden; single platform, Snapdragon reference design only), but not the strong ones. Android did it the other way, so I'm not surprised at all.
> it turned out that having a world class OS was more important than perhaps anyone at the time knew.
Microsoft never fully committed to a phone OS because workable tablets might have demolished the desktop market.
The problem was that the "Windows XP Tablet Edition" was utter crap. A desktop OS and its mentality shoved into tablet form with modifications that fell way short of what was actually needed for that form. The iPad was the first tablet that got it right - and feel like a tablet meant for tablety things, as opposed to a PC without a keyboard and crippled in all sorts of ways.
If the cruise ship sinks, the yacht might be a great help - or it might get pulled down with it.
Luckily, the cruise ship doesn't show any signs of sinking at the moment.
As for the "edge of bankruptcy" was due to a series of bad business decisions, not the Mac line itself. If anything, the Mac line and Mac customers was the only thing they still had going for them. It was the OS that was old and tired in an era of good enough Windows, Apple wasting resources in Copland, Newton and other endeavors without results, the blunder of allowing clones that ate into their Mac lunch, and so on.
edit: 2023!
> It's just choice.
Apparently, one of the reasons for its lack of adoption, was lack of a native macOS GUI – instead it ran under X11. https://community.sw.siemens.com/s/question/0D54O000061xMQgS...
If Apple really cared about it, I'm sure they could have worked out some deal with Siemens to keep the macOS edition viable, even improve its UI to be native and more usable. It would require them to move away from viewing themselves just as a customer, into more seeing it as an investment into their own platform (making it viable for more use cases.) Sad to say, Apple doesn't seem to think like that much (or maybe some people there do, but not the people whose opinions really count)
For CAx in general it's difficult to say, because there's such a huge variety of software and every little niche has its own highly specialized (and usually very expensive) tools. I've never seen any of these support anything other than Windows and/or Linux. Some of them have Win32 GUIs and Linux is headless only (for running simulations), some are fully cross platform, some are Linux-only.
Most CAE engineers around here work on Linux, mainly because they prefer how easy it is to script and automate things and because the headless HPC/batch environment is of course also Linux. Meanwhile CAD and design only happens on Windows because CATIA is Windows only, like the various Autodesk tools used for CAID. From the late 80s to the early 2000s this place had a ton of UNIX workstations (SGI, HP, Sun) and even UNIX clusters (IRIX and SUPER-UX among others, the latter having virtually no representation on the internet today). There's also still IBM AIX systems around, as well as IBM mainframes. Not my department though.
Are you saying some people still use IBM mainframes for CAD/CAE/etc applications? If yes, that's unexpected yet intriguing information, and I'd love to know more.
Although maybe you were just stating the obvious that IBM mainframes survive in general even if no longer in this particular domain.
Suppose something like CATIA was available on Mac and offered lets say a performance benefit. Would you consider Mac? Or would you for example still need a whole other set of tools to be available as well for it to be even possible?
I haven't had the time or the desire to take it for a spin yet, but I'd love to be able to compile FPGA designs on a nicely configured and quiet M2 system.
And without external DRAM, they’ll never be able to provide as much memory as a low-end Intel chip in the Mac mini at the same price point. In-package DRAM is better performing but it changes the price/perf graph.
Apple simply optimized their silicon packages for 90% (or more?) of their users over the outliers.
The bigger is how much LPDDR5 memory they can hook up to their 512bit memory controller.
(No idea if it's so, but DRAM is such an analog HF beast that a controller needs training before it can work with a particular chip, in a particular slot on a particular system board. It's automatic but it's there, and it must have limits.)
Surely that would require architectural changes and the software would need to be adapted but since these machines are niche, so is the software and Apple can sponsor the development of the few software people are using on that kind of machines. Apple is sponsoring Blender for example.
I think the size of the step between a and b is becoming too small to be viable: not enough extra performance, and too high price compared to step c.
Dedicated HW boxes connected using Ethernet/Thunderbolt/USB-C have the great advantage that you are no longer bound to whatever Apple offers this year. You will just use the Mac/Windows as a "advanced graphical terminal".
Same for hardware developers. Looking back at the past 10 years of MacPro shows great uncertainties in what Apple offers going forward. Why would you spend time making a PCI card for MacPro when Apple so easily abandons support for PCI with the TrashCan while you could have a reliable market with a thunderbolt/ethernet accessory?
IMO the M1 Max studio body is unnecessary too, it's only really needed for the M1 Ultra, which would lead to barely anyone buying the mac studio form factor.
The Mac Studio must be decimating Mac Pro sales, it must fit what 70-80% of Mac Pro users wanted. It's going to be incredible hard to Apple to justify development of an extended M architecture for a product used by a tiny fraction of their customers.
It wouldn't surprise me if they didn't press go on full Mac Pro development until they saw how the market reacted to the Studio.
One possibility is they reintroduce the XServe (and maybe XServe Cloud) with the larger expandable architecture to help justify its development.
The only other option is expanding the Mac Studios capability with external "stackable" peripherals, such as eGPUs.
I've been thinking for a while they'll bring XServe back (or the Mac Pro will be rackable) because they'll save a TON of money.
1. Lower power usage in their own data centres - that's gotta be huge.
2. Buying processors from themselves instead of from Intel.
Hopefully it will help get whatever chip they're going to use in the Mac Pro to some kind of production number that means economies of scale come into play.
Then they finally released the 2019 Mac Pro at a cost that I thought was prohibitive to all but the highest end professionals. I don’t remember previous Mac Pros being targeted this exclusively; heck, somebody in my freshman dorm in 2006 had the cheese grater Pro at the time for music production.
When they unveiled the Studio, I thought that they were replacing the Pro line with that product…but then they confirmed the Pro is still coming. I really do not know what differentiation there will be between the Studio and the new Pro. I’m curious to see what they do with expansion and dedicated GPUs, but considering they don’t play ball with Nvidia anymore it’s probably still going to be a niche product
With that change, laptops have become the computer market. Look at Apple's iMac offering; they no longer have multiple models and the insides are now laptop parts. The Mac Studio is simply the desktop computer that gets the Macbook Pro chips. Since they don't want to engineer a cooling solution behind a monitor every time, we get a simple box with a ginormous fan.
The Mac Pro was supposed to be the computer with the truly desktop-only ARM chip. Even that seems cancelled now. I don't know what Apple is planning.
«Laptop parts» connect M1 CPU clusters to RAM via a 512-bit wide (M1 Max) and a 1024-bit wide (M1 Ultra) memory bus yielding 100 Gb/sec (per a CPU core), 200 Gb/sec (per a CPU cluster) and 400 (M1 Max)/800 (M1 Ultra) Gb/sec cumulative data transfer speeds. GPU and ANE cores access the main memory via the same wide memory bus. L1 caches (I+D) are 128kB+64 kb (efficiency cores) and 192+128 kb (performance); L2 caches are 24 Mb (M1 Max) and 48 Mb (M1 Ultra), and L3 cache is 48 Mb (Max) and 96 Mb (Ultra). Auxiliary tasks (disk encryption, video encoding/decoding etc) are off-loaded to dedicated processing units which is reminiscent to that a typical mainframe architecture with channel processors and channel programmes rather than that of a laptop.
Most «hi-end» servers today do not come anywhere close to such configurations leave alone the form factor and the net system efficiency. 512-bit wide memory bus is currently only found inside hi-end GPU cards, and the compute off-loading / accelerating units are an add-on feature for a server that may or may not be available.
Only POWER10 and Telum systems from IBM offer similar (albeit far more scalable and feature rich) designs and at a completely different price scale.
AMD EPYC CPUs just went from 512-bit DDR4 to 768-bit DDR5, per socket.
I ought to have used «typically» rather than «currently» as wide memory buses were coming and going more mainstream in the server space, whilst not yet being widely available today. The hardware transition cycle will take a few years, and as of today «currently» still holds true, semantically and factually, as cloud platforms and data centres run nearly entirely exclusively on Xeon something Lake and EPYC 3rd Gen CPU's, and neither has such a wide memory bus.
> … per socket.
Does that mean that 2x socketable EPYC CPU's (say, 512-bit wide each) will be able to access RAM via a 1x 1024-bit wide memory bus? If yes, how will access to RAM work, i.e. access multiplexing – memory banks? time interleaved access? something else?
No. All recent dual-socket platforms are NUMA, meaning a CPU only has fast access to its local memory, and accessing memory attached to the other CPU socket requires going through a slower link between sockets.
NUMA-aware operating systems and applications try to avoid allocating memory on a different NUMA node than the processor cores that will be using that memory. On AMD EPYC, the internal structure of the IO die (memory controller) means it is sometimes advantageous to enable the option for the system to report more than one NUMA node per socket.
If I read it correctly, 512- and 768-bit wide memory bus is available to access the socketed CPU's local memory only (which is relatively small), and access to the shared memory (which is large to very large) seems to be still via a switched Infinity Fabric memory controller.
How wide is the Infinity Fabric nowadays? I can't easily find the conclusive evidence.
Let's forget the whole Apple Silicon thing, and look at 2020. The difference in single core is minuscule, but the big difference is multicore. Unless you need those extra cores, those computers are equivalent.
Now compare a 2010 Mac Pro and the 2010-2012 consumer laptops. Single core is 100%, multi-core is at least 400%. That's a big delta and that's only 2010.
If you go back into the past, it would surely ebb and flow but surely stay in that zone. I'd expect a laptop to be half the speed of a big pro desktop all the way to the prehistoric times when all computers were beige and laptops charged extra to get rid of screen ghosting.
And the more you move into the future the more a laptop is fine.
Well, even before the rise of covid-induced work-from-home, Apple has been shepherding almost all of their users onto laptops.
That college student's 2006 Mac Pro might have had 1GB RAM [1] - but these days, Apple has a laptop for users wanting 64GB RAM. And it isn't a suitcase-sized mobile workstation either.
People wanting more power than that for things like ML and scientific computing will often be shifting their workload to the cloud or a cluster anyway - it's only people doing interactive tasks like CAD and high-end video editing who are buying the Mac Pro.
I develop bioinformatics methods and software, and I see significant productivity gains when I can run tests with full datasets locally. So far I've been using an iMac with 128 GB RAM, but I'd like to upgrade to something with 256 GB or even 512 GB. As the Mac Studio is pretty underwhelming from my perspective, the options are basically a hypothetical Mac Pro and a Linux box.
most offices I’ve been around have internet that tops out no more than a little faster than a decent home connection, but have dozens of people sharing them.
[1] https://www.ecseq.com/blog/2020/Whole-Exome-Analysis-on-a-Ma...
Whatever it is they're struggling to get the performance they need for a Mac Pro level machine.
I'd imagine part of it is going to be that they'll want to continue using Nvidia or AMD GPU's to allow for higher performance. That can't be an easy or simple thing to get working with their custom ARM based architecture.
Maybe, but people are always coming and going on any team. This whole story felt like a tempest in a teapot type of thing.
IMO, given the typical lead time of hardware, we're finally seeing Apple have issues stemming from covid/supply chain breakdowns.
"The ViewFinity S9 (Model Number: S90PC) is Samsung’s newest addition to its monitor lineup. It features a 5K 27” screen for the first time and is optimized for creative professionals such as graphic designers and photographers. Its 5,120 x 2,880 resolution combined with its wide color gamut of 99% DCI-P3, provides crisp and true-to-form details, and its average Delta E ≦ 2[1] color accuracy also produces clear and precise color representation, even in complicated or nuanced visual environments."
https://news.samsung.com/us/samsung-electronics-unveils-new-...
I really like the 30" Ultrasharp, and would consider a 32" as long as the resolution, connectivity, pixel density, and panel tech are in place.
Thunderbolt 3 has been around for over five years on mainstream Mac (and Windows) machines, so what’s the excuse again?
I can count the number of monitors in the same resolution space (or higher) on one hand! ;) There’s the insanely priced XDR 6K, there’s the not much better priced Dell 8K, then there’s the LG 5K that’s cheap plastic build quality and hasn’t been refreshed in 5 years. Anyone else? … Bueler?
Most of the M-series chips are shared across multiple products, but the Mac Pro seems to be a clear case where using a powerful off-the-shelf Intel chip makes more business sense right now.
I don’t see Apple doing it simply as a vanity project, so there will have to be some innovation or new product line coming if an Apple Silicon Mac Pro is to ever exist.
It will simplify MacOS when they end Intel support.
Not only a cost to Apple, but also a cost to 3rd party Mac developers, who would be even less likely to have the resources to support the Mac Pro for its own sake.
Apple dropped PowerPC support in Mac OS X (10.6) 3.5 years after the Intel transition started, and they dropped Rosetta support in Mac OS X (10.7) 2 years after that.
IIRC the last PPC Mac hardware was discontinued very soon after the Intel transition started.
Anyway, no, not for Apple either, because Mac Pro unit sales are already low, and trying to make that one Mac model cover the cost of all macOS Intel support will make the already expensive and low volume Mac Pro even more expensive, and thus even lower volume, making it harder to cover the cost.
The difference would be supporting Intel publicly - with third-party developers. As the target for Mac Pros won't just be running a load of electron apps.
The question is how they will pay for a high-end chip when they no longer can rely on Intel to spread out the development cost among many customers.
Frustrating those people likely comes with a risk of losing high value customers.
I think it’s similar to the model a lot of game consoles have. They make little to negative profit on the console itself during the first couple years, but make it up on software and other benefits of hardware lock-in.
In the time after the trashcan Mac Pro they lost a significant portion of the professional market. And while that market is relatively small, it's highly influencial.
By the time Final Cut Pro X was unveiled Apple could boast that 60% of Hollywood movies are edited on a Mac. Now? Not so much. And this probably extends into other areas, too.
There was no version 9, nor version 8. ;)
They can probably heavily base a more powerful chip on their existing designs but with more cores & cache, so design should be relatively cheap. Manufacturing the masks etc. is probably around $10m which is probably fine given how expensive Mac Pros are (and it would simplify their software if they can ditch Intel entirely).
Somebody buying a mac pro would probably expect support for discrete and expandable ram.
Maybe it needs more PCIe lanes but again that probably doesn't need much design work.
Which is the question I have been asking since 2015. From a cost prospective it never made any sense. Presumably they will wait until Mac Studio becoming powerful enough to fill the gap.
>but the Mac Pro seems to be a clear case where using a powerful off-the-shelf Intel chip makes more business sense right now.
It surely does. Except it will be the only Mac in the lineup without Metal Support and Apple's increasingly important NPU for Machine Learning features. I dont see it as a good trade off.
To me the only reason they still make products like the Mac Pro is for brand reputation and face-saving. It’s like when a big auto maker sponsors a formula 1 racing team. There’s no profit in formula 1 (for the auto makers). It’s just a giant money pit!
If I'm starting out today and looking to choose a PC or Mac ecosystem, if I see that the Mac has an upper bound that's lower than the PC, then in my starry-eyed optimism as a newbie I'll think I might as well start as the PC since one day when I'm super successful and I need more power I'm going to have to change platforms anyway.
Even if none of those users will ever end up using any more power than the Mac Studio, if they're convinced they will then they will not choose your ecosystem.
Some caveats:
* maybe not true if you go back to the introduction of metal but idk
* metal feature support varies by hardware
https://developer.apple.com/metal/Metal-Feature-Set-Tables.p...
I’m not sure how to respond as a result since it’s hard to discuss in good faith.
V1 of their comment simply asked what “metal” evokes in one’s mind.
V2 then posted a link to the Metal page saying it lets you directly use your GPU hardware.
V3 is now pointing out that metal feature support varies by hardware which I already called out, but they’re positing as a contradiction somehow.
As such I’ve removed my reply because it makes no sense based on their frequent changes and left this one instead.
As for your latest update, I had already mentioned that in my comment as a caveat, though I hadn’t shared the link which is always great to have.
The two tricky things I'd anticipate are:
What does it look like to hardware? Different CPUs speak different protocols to maintain a cache coherent view of memory. So they couldn't easily share memory in a completely natural way - but GPUs manage to share memory between host and GPU, so it can be done (just not necessarily as conveniently to software).
What does it look like to software? Lots of weird options here. Most convenient for software compatibility is it the Intel is also somehow running MacOS (or appears to be). I guess you could do that - it'd be like having a two node cluster in one machine. But that's not as convenient as being able to distribute computation between compatible CPUs.
None of that is a complete technical blocker to a company that controls their stack, I'd just guess it would make it too inconvenient / expensive for them to do.
That said - as an intern, years ago, I once had access to a piece of prototype hardware that had an x86 chip in one motherboard CPU socket and a non-x86 chip in the other (with various adaptation hardware to allow them to play nice). I think that was more a proof of concept than anything else, not necessarily viable without other hardware changes. Interesting nonetheless.
I saw a lot of new-in-box examples cheap on eBay prior to Sun's implosion, so I suspect they weren't particularly popular and were primarily targeted at a few important customers that had some box to tick.
Eg: https://everymac.com/systems/apple/mac_quadra/specs/mac_quad...
They are not starting from scratch. The Apple Silicon MX chips all share a common base, and are related to the AX chips. When every product Apple sells is using some version of the same family of chips, a lot of the design cost is spread across all Apple devices.
Then it's just the incremental design cost that must be recouped for things like (speculated) off die RAM.
That is until you run the calculation when the cost of doing it are in the tens of thousands per unit. Suddenly it doesn't make much sense.
Sure, Threadripper makes economical sense for AMD only when "subsidised" by EPYC/Ryzen sales and the shared chiplet architecture. But Apple has deep enough pockets for halo products. They are allegedly repurposing some of the binned/sub-par silicon into internal server hardware anyway.
In their own datacentres?
AWS is already doing so with Graviton.
It has nothing to do with macOS.
Graviton CPUs are not for sale. If Apple wants a million ARM cores to run iCloud, they are going to have to get TSMC to fab them.
(Often, when one application crashes, a couple others follow not long after.)
Your issues are not normal. For example, here on my Studio, the uptime is 9 days, with Sublime Text started from the beginning and used quite heavily during most days. It has 5 windows with 5 to 10 tabs in each with no issue whatsoever. My experience with my previous (Intel) Macs was about the same; I haven't seen visible effects of the ISA change besides not hearing any fan anymore (and a speed bump, but I tend to use my computers for a very long time, so there is always a speed boost when I get a new one).
If it is under warranty, it's probably worth having a chat with Apple support. Otherwise, I would recommend trying to diagnose hardware failures (particularly RAM; beyond 8GB it should not swap regularly-used applications).
MacOS is rightly criticised for becoming less robust in several ways, but that is mostly fancy new features. And Time Machine. For some reason they cannot seem to get it perfect, which is a shame because it is the best backup solution I have used yet. The core OS is quite reliable and really should not crash things like that.
[edit] There is a wealth of information about AS Macs here, the site is well worth a bookmark: https://eclecticlight.co/mac-troubleshooting-summary/
On qBittorrent specifically the application can often "feel" like it is hanging because it brings up modal windows behind the main window, meaning you need to drag the main window away to interact with the modal and the main window will be unresponsive in that time. This may make it feel like it is hanging, when really it's just down to a non-native macOS application not using the correct APIs to display modal sheets
I do occasionally get other application crashes, maybe a couple per month, usually it's down to the application doing something it shouldn't. My current macOS uptime is 28 days, which is when I last installed an update that required a restart
To be fair, MacOS itself doesn't seem to have issues. There seems to be no upper limit to possible uptime. It's just applications.
But without knowing the machine and OS version that you have my data point is perhaps not very helpful.
I would definitely take it in to be looked at. That should not be happening. Crash logs will hopefully point a finger at the problem.
Have you tried reinstalling everything from scratch? If yes, then it is very likely a HW issue. If others suggest a better method for checking where the fault lies, other than reinstalling, then follow their advice.
On the other hand Microsoft's surface lineup makes no sense and is super confusing. That's true for Samsung as well. Even if these companies had the engineering capacity to make chips like M1; I doubt they would attract even a fraction of the media attention enjoyed by Apple.
They've kinda gone that way again with sticking the "Pro" and "Max" terms onto everything, especially the iPhone models.
And with Mac Studio, which brings us the next question. What are the Market Mac Studio couldn't compete with Mac Pro. If it is not powerful enough now, will possibly another doubling of LPDDR6, that is 512GB with 1.5TB/s of bandwidth with M3 Ultra be enough?
The only market I could think of right now is Audio Professionals with specific Plug in Sound Card. Could that market not be served if Apple had made a few more Thunderbolt 4 port available?
Windows is already in many studios due to VR or Virtual Production work, since Unreal Engine kind of sucks on anything that isn’t Windows, and the same goes for VR.
However there are a few areas that Macs are used heavily and where I don’t think Apple need to be worried:
- artists that need to use Adobe products or anything non Linux. Macs offer better Linux like workflows for pipelines since it’s Unix like, so is favoured for pipeline integration and macOS has much better colour handling than windows. So you’ll see in house concept artists, matte painters and the like on macs. VFX editors tend to be on macs for the same reason. Those are the folks that a Mac Pro would be handy for.
- dailies review boxes. Macs again are better for playback and the arm macs are so quiet they’re perfect for a review room. More RAM is always welcome here.
- coordinators and producers are almost always on Mac laptops. They’re unparalleled for battery life in a compact body. Plus since macs need to be supported for the above two use cases, it makes sense to use them as well.
- conform machines to handle media conversions often are macs. Though this depends greatly on what formats are being used
If you stick to apps that are specifically colour managed, and manage your own screen profiles, it’s fine, but compare it to the Mac where:
* you get much better EDR/HDR support at the OS level, so it tries to handle perceptual EDR ranges when apps switch to EDR.
* the OS itself has better management for colour profiles for the screen (also better refresh rate management when it comes to common film rates)
* Macs come with P3 colour handling for their first party displays https://support.apple.com/en-ca/HT210435 , and while you can colour manage a windows/Linux machine post facto, it’s nice to have a system come with a good baseline that reflects their screen properties. For example, coordinators don’t need colour accuracy BUT it does help when their laptops present it well. Most desktop users would use an Eizo instead.
* safari does a better job than chrome/Firefox at handling embedded image colour metadata , and also managing EDR colour.
* similarly Preview, QuickTime and Finder are better at respecting embedded colour profiles and EDR where applicable than Windows Media Viewer and Explorer.
Most of this can of course be replicated on Windows, but you have to go out of your way to do so. Plus you’d still lack the macs EDR handling that is great for review sessions etc…
macOS doesn’t cover everything of course. Images with custom LUTs still need their own viewer apps, but I can’t emphasize how nice it is to have much more trust in what you’re seeing and reducing the mental friction while doing a lot of your ancillary work outside specific apps. It’s not perfect, but so much better then Windows where it’s a crapshoot what colour you’re going to see depending on how you view it.
I don't see how they can solve this either.
I’m not sure I agree that a TBDR is bad for high poly counts. There’s nothing inherent to the architecture that I can think of that would limit it for pure polygon work. Usually the biggest hold up is full screen effects that need to work on the entire buffer but those are rare for VFX.
Will DRAM really scale to that extent at the device fabrication level, though? We're already seeing limits to scaling with SRAM.
I am assuming you mean the transition from TSMC N5 to N3. Which was expected. N2 with GAA will get some density improvement.
Originally, TSMC projected 3nm risk production to start at the tail end of 2021, so Apple would have had chips for multiple product lines that were designed to be fabbed at 3nm in 2022.
Or is there a more technical hurdle, e.g. that supporting external Dram and third party GPU would require more fundamental changes than merely building a new SoC?
So the only thing that makes sense is to either make no Mac pro tower workstation with discrete GPUs at all, or doing what they did last time, using third party cards.
I doubt they will release a M2 Pro Mini. They will instead update the Mac Studio. Conversely, if they make a M2 Pro in the current Mini form factor, they'll likely use that as an excuse to redesign the regular M2 Mini to be smaller (Mac Micro) and more compatible with DC-power data centers.
What is the state of the art, do all of you simply buy a top-notch Macbook Pro and work on it all day?
It's the best machine I've ever had. But then again I've only owned laptops since 1998. Still keep my 2014 Macbook Air on the desk just as a conferencing / music box.
[edit] I also only work from my home office when I have to... the rest of the time I'll be working on the porch, at the kitchen table, out at a cafe, etc.
Almost always with 2x 27” 4K Displays.
Nothing fancy, just the Dell range and whatever I can get a good price on.
I’d prefer 24” but you can’t get them at a sensible price. I’d use Apple displays but they are out of my budget range!
I tried 32” 4K displays, but they are too big for me!
Almost never gets used as a laptop. It just moves from desk to desk.
I get the 16” as when I do need to work directly on the MacBook, I want as much space as I can get!
It’s more convenient than a Mac Mini, since I have the same machine everywhere.
Or they’ll save the Nano name for a iPhone-sized Mac.
The difference could be in expandability (PCIe slots) and RAM if they are able to either move the RAM off-die or have a NUMA architecture.
Interesting to speculate though. So far they also haven't launched a multi-socket product, maybe that's what the Mac Pro could launch with.
Developers either target apple silicon and ignore intel, ignore apple silicon and rely on rosetta2 for as long as that's here, or invest in both to optimize performance and just eat all the extra work both require? I'm not a mac user, is there such a thing as making a app targeting apple silicon work on a x86 machine?
If someone is looking to add their application to Mac it really seems like a bad time to do it, wait a year and more apple silicon will be out there and you could reasonably target just that and be OK. Apple prolonging the transition is not good for anybody, but I guess we can blame this on supply chain.
They currently support windows, redhat, and just added Ubuntu. Very much similar version of the software runs on their "native" jace 8000 (PLC lite) hardware, which is ARM under it all so you would think they already have experience in that area.
I think you are probably still incorrectly assuming that the CPU architecture and instruction set have any relevance. What you've described sounds like the kind of software that is a portability challenge due to being tied to specific operating systems and their particular APIs for interacting with peripherals. It doesn't sound like anything that should much more than a recompile to target a new CPU architecture, provided that the application has already been ported to the operating system that will be used on that architecture. Certainly nothing you've described would require hand-tuned assembly code to reach adequate performance on a decade-old laptop.
Just like sports, a team can’t stay at the top forever. Now that Apple has lost many of their Avenger members in their chip division, they need to be humble and go back to the drawing board. AMD and Qualcomm are getting closer to their performance per watt.
The M2 showed that the M1 architecture is done. They need to build something new up from the ground up but that’s going to challenge. Let’s see if they can pull it off.
How so?
In some respects it didn't work out well as an independent, as it was later acquired by (absorbed into) Qualcomm, which is already a large processor design company. But if it was cash the founders wanted, they got themselves a good exit.
In fact the only recent Apple employee other than Jobs or Cook that I have ever heard about is Jony Ive.
Threadripper Pro exists and as far as I can tell, certain industries are practically begging AMD to keep updating it. A system with 32-core TR Pro and a few hundred gigs of RAM is easily $10k. With a 64-core CPU, you are probably closer to $15k. If a Mac Pro can give you 40-60 cores, 128-512GB of RAM plus a powerful GPU for $10-20k, its certainly niche, but not dead on arrival.
I think the main issue is that while AMD can "justify" TR Pro by having disaggregated compute and IO chiplets, Apple does not. AMD can just "print" as many CPU cores as it likes (within the power budget), and then slap on a different IO tile depending on what product it's making.
Currently, Apple isn't doing that. Furthermore, they put themselves into an even bigger corner by gluing their CPU and GPU onto one die. Most professional tasks are fine with a monster CPU and no/little GPU, or a monster GPU and an ok CPU. Adding PCIe with support for GPUs would help with that. For memory heavy tasks, you could add CXL memory expansions. But that then adds a new problem: what dies would Apple fab for this product?
With M1, there were essentially 2-3 dies: The power efficient M1, the M1 Max/Ultra with the interposer, and M1 Pro (there were rumours some Pros are just failed Max-es with a part of the chip cut off, but afaik nobody was able to verify this, so 3 dies it is). For M2, we have only seen the power efficient die so far. If M2 Extreme would be 4x M2 Max, then the previous strategy could work. But if Apple wants PCIe or CXL memory modules on M2 Extreme, they have to put it on all of their Max (and maybe Pro) dies as well. Even though these will be in laptops/all-in-ones that won't use any of that. In other words, a ton of analog circuitry that's a complete waste of silicon. The only other option is to fab special dies for M2 Extreme, which might have been the plan, and what probably made it a very bad ROI.
It's kind of like Sapphire Rapids: SR is certainly fast for what it's doing, but it's super expensive compared to EPYC and less scalable exactly because you need different dies to implement different SKUs (SR is monolithic up to afaik ~20 cores and the "chiplets" only appear above that, plus the IO you get depends on the chiplets). And SR doesn't even include a GPU...
Finally, Apple's GPUs don't appear to be really scaling as great as they were hoping. I have yet to see a real-world M1 Ultra review that wouldn't end on a "it's +20-50% more performance for +100% price" note. The CPU cluster seems to be doing much better. But by tying four of them together for M2 Extreme, people who need CPU would have to pay for the whole thing, and people who need GPU would be probably left disappointed.
My personal "best bet" would be to just keep releasing a power efficient "consumer" die, but disaggregate everything on the "pro level": A single CPU-heavy "Pro" die with a bit of GPU and a bit of IO, then a GPU-heavy die that slaps onto that to make the Max, and then an IO-only die for the Mac Pro mixed and matched with a combination of CPU and GPU dies. I wouldn't be shocked if this was the plan for M3/M4, the question is if the packaging technology will be there in time (and economical; wink wink Intel).
My M1 MacBook Pro can run full throttle on all 8 "performance" CPU cores for an hour without the fans ever spinning up.
This tells me that Apple could fit 32 performance cores in a laptop CPU, which is simply not possible with x86 (unless you severely compromise on the definition of "portable").
I'm not too sure about that. The current 20-core CPU from Apple is inside the Mac Studio, and it has quite a monster of an air cooler. It's obviously going to be much harder to put that (or something more powerful) inside a laptop, even with all the chip improvements...
Something’s not adding up here, though. 8 cores = “tiny fans basically always off” vs 20 cores = “two huge fans”. Either those two huge fans are running at a very low RPM, or the fans in the laptop are almost as efficient as the huge ones. I highly doubt the latter, so I assume the former is the case.
MBP 13: 4 performance cores + 4 efficiency cores
Mac Studio: 16 performance cores + 4 efficiency cores
Not to mention that the more cores you have, the bigger the overhead of e.g. synchronizing between them becomes
I predict that Mac sales will sag and that this will lead Apple to slash more and more models, eventually undermining the legitimacy of the platform. Most users will switch to Linux on Intel laptops from various other vendors.
Meanwhile, Mac sales are up and the current crop of M1/M2 machines get top review marks.
Who exactly will switch to Linux on an intel laptop outside of a fraction of software developers/IT professionals?
It would be helpful if you have facts or data to back your statements up.
The amount of users using windows on a mac is a negligible fraction of the customers. I personnally don't know a single person using a mac either for personnal or professionnal use that is firing up a windows OS on it. People loving the windows platform or who need to run some apps on it are choosing computers from other brands.
Booting windows on a mac appeal only to some developers who need both platform but they can easily remote connect to a windows if they need to.
c/famously/anonymously