Good luck finding a 5K external display that works properly on a Mac without compromise of some sort.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
'iMac Pro' Coming in 2022 With M1 Pro/Max Chips, 27-Inch Mini-LED Display, MacBook Pro Ports and More
- Thread starter MacRumors
- Start date
- Sort by reaction score
I think SDJim's point is probably more applicable to the laptops, where you can't get a large screen size without getting heavy-duty processing power. Now that the large MBP has transitioned to being workstation-class, Apple doesn't currently have a consumer-grade laptop with a large screen. I.e., I think what they need is a 15" or 16" MacBook Air.
As for the Mini+5k, yes, you can do this. But, as a minor point, as I'm sure you know, you can't do this if you want both your computer and display to be Apple. They make the Mini, sure. But the only external monitor they've produced in the past several years is the $5000 XDR. The only way I know of to accomplish what the OP wants is to buy a Mini + LG 27" 5k (https://www.apple.com/shop/product/HMUB2LL/A/lg-ultrafine-5k-display?afid=p238|sCyj0Fx3O-dc_mtid_1870765e38482_pcrid_500452391010_pgrid_43272004728_&cid=aos-us-kwgo-pla-btb-3pp--slid---product-HMUB2LL/A). And once you've done this, you might as well get a 27" iMac.
Last edited:
The 27" iMac Pro should have a much higher ability to dissipate thermal energy than the 16" MBP. Thus, even if they use the same chips, they could potentially clock the CPU higher (and maybe this could apply to the GPU as well?).
But I've no idea whether they will. It may be those chips simply aren't designed to run faster than they already are—it's possible they could become unstable at higher clocks and/or that the efficiency becomes very poor.
Another possibility is they offer the same chip design, at the same clocks, but with higher GPU and/or CPU core count options, which would result in higher thermal output.
OTOH, they might offer no more processing power in the iMac Pro than they currently do in the MBP, so they can make it super-thin like the new small iMac (I hope they don't go this route).
Last edited:
No. The only thing they will maybe do in addition to M1Pro and Max is use two M1 Max SoCs in a package.
Your post says it all.
Nota compelling reason to ditch the i9 in a non-mobile setting, barely faster (10%) than AMDs best MOBILE chip.
And at what cost? 32GB and 1TB with the Pro costs $3000 in a MBP.
The M1 is a compelling mobile chip evolved from a phone chip powerful enough to coordinate a score of concurrent Apollo missions in real time.
But one expects a PRO desktop chip to try to be a super computer, not a power efficient mobile chip shoved in a PRO enclosure.
Don’t need something so hot it requires liquid cooling like the G5, but at minimum take advantage of a desktop form factor and heat things up. Making the iMac skinny was an excuse not to have to justify the same speed as a fanless MBA.
Edit: and don’t forget lack of discrete graphics. On mobile, fine. In a desktop? Fugedaboudit.
Last edited:
To keep decent pixel density only offer i’m aware was LG’s. XDR 32”6k is costly overkill. So hopefully we will get a product from Apple.
After years of crappy graphics, and the full adoption of new SoC concept, it seems very very unlikely we will see support for 3rd party discrete graphics (not even mention nvidia) except for MacPro.
I see things from different perspective: Although I’m expecting a performance bump in the SoC of new iMac, I think we’re now at a stage where mobile chips caught up desktop performance, so gladdly, this means we don’t need to compromise when going mobile.
I'd really like to see Apple take advantage of a desktop's greater thermals to offer a more powerful class of chip than what's available in the MBP's. While it may be only a minority that need to use more than 8 performance cores simultaneously, many could still benefit from significantly higher single-core speeds. If Apple could deliver that at the cost of power consumption that's too great for laptops, but fine for desktops, it would be unfortunate if they did not do so.
Plus it looks like Apple's marketing is now about having the fastest CPU in each class. They've achieved that for phones, tablets, and laptops. But if they don't give the iMac Pro chips that are more powerful than what's in the MBP's, they'll be ceding that crown back to Intel and AMD for desktops. Intel's top-end Alder Lake desktop chip, without overclocking, may have a single-core GB speed that approaches 2,000.
Plus it looks like Apple's marketing is now about having the fastest CPU in each class. They've achieved that for phones, tablets, and laptops. But if they don't give the iMac Pro chips that are more powerful than what's in the MBP's, they'll be ceding that crown back to Intel and AMD for desktops. Intel's top-end Alder Lake desktop chip, without overclocking, may have a single-core GB speed that approaches 2,000.
The way I see this, this level of performance cannot scale. The workstation segment, which the iMac Pro and Mac Pro sits in is all about scaling out the core counts. Scaling up frequencies will make scaling out difficult as cooling it sufficiently without it melting will be a lot more difficult with many cores pumping out heat.
I suspect the Mac Pro will use the A15 cores for better efficiency and probably going up slightly in frequency. If rumours are accurate, it will have multiple Max SoCs stitched together for up to 40 CPU cores.
Sure, and that's why the Xeons have poorer single-core performance than i9's.
But:
Apple has a large class of desktop (iMac) users that wouldn't benefit from a lot of cores, but would benefit from very high single-core speeds, i.e., from performance more like an Alder Lake with 8 performance cores than a Xeon with 28. Indeed, I'd say the overwhelming majority of Apple desktop users fall into this category.
Given this, wouldn't it make sense for Apple to provide a chip for this large customer group that, while it might not scale well to high core counts, would give extraordinary single-core performance?
It's not as if they'd be abandoning those who'd like a workstation-class iMac by giving them what I described above instead of MBP chips, since if the new iMac is going to use the M1 Pro/Max chips, it won't be workstation class anyways. Those chips are workstation-class for laptops, but they're not workstation-class for desktops (unless they use multiples of them, which doesn't correspond to the current iMac rumors). So putting M1 Pro/Max chips in the large iMac doesn't help either of these user groups.
Last edited:
IMHO, the M1 already satisfy the majority of iMac users' needs at the moment. Apple will never be able to cater to (nor do I think they want to) everyone's needs.
And I don't think Apple will chase after the performance crown. They will just hit the target they are aiming for and aim higher next round, and their targets always revolve around the products they want to build, not around any raw performance metric.
Alder Lake may look good on paper, but I suspect real world performance will likely disappoint. The "open" PC architecture just have too many bottlenecks (e.g. PCIe, south bridges with more PCIe, SATA, WiFi, etc.) all vying for memory bandwidth, which is 77GB/s 4-channels DDR5-4800. This is roughly similar to the M1, and Anandtech found that a single M1 core could consume all that memory bandwidth. I'll wait for Anandtech to perform a more in-depth analysis of Alder Lake to see if there's any improvement there.
Edit: corrected memory bandwidth figure.
Last edited:
It's not about catering to everyone's needs, it's about catering to the needs of a large block of Apple's customers. If you think the M1 is fine for everyone that buys an iMac, then your position is that there's no need for anything but the M1 in the iMac. But I don't think anyone, including Apple, believes that--at the very least, the large iMac (aka iMac Pro) will have the Pro/Max models.
Alternately, if you admit to the fact that the iMac Pro does need more powerful chips than the M1, then it becomes legitimate to discuss what those chips should be, and your "M1 already satisfies..." argument no longer applies.
I.e., you can't have it both ways—you can't both admit the iMac Pro needs more powerful chips than the M1, but at the same time dismiss any argument about what the nature of those chips should be by saying "the M1 is enough".
Last edited:
I'm not saying that iMac users don't need more powerful chips. That's why the M1 Pro and M1 Max exists.
I don't think that Apple need to beat out Alder Lake to produce a better 27" iMac/Pro replacement. The M1 core is already good enough. It's just having more of them in a single SoC. Next year we will likely see the A15 or maybe A16 CPU cores in the new M SoCs.
My take is that Apple will settle down with 3 categories of Macs: good, better, best, if you will.
The M1 covers the 'good' for now. Q1 2022 will likely see M2.
Apple just started on the 'better' category, and the 27" iMac replacement will be next in line for the 'better'.
The 'best' will be the Mac Pro, and possibly the iMac Pro replacement. Maybe we'll be surprised with a 'best' category MacBook Pro? I suspect this category will have later generation CPU core architecture than the M1, which we will then get the single core CPU performance uplift.
I don't think Apple now wants to constraint themselves by artificially segmenting performance between desktop and mobile. It's just performance classes to them because AS allows them to do that.
Why do you want discrete graphics which are most definitely on a slower bus and are more expensive and might hinder hardware development in the future because of contract issues and release cycle asynchronicity. Also if Apple is able to scale their GPUs on their SoCs further like they did until now there should be no problem performance wise.
I believe Apple will be make a M1 Pro Desktop model of the iMac, but I have my doubts that they will name it the iMac Pro. Though I could be wrong.
Why do we need discrete graphics if we can get near 3080 Mobile performance from a SoC?
I'm pretty sure Apple can get 40 Tflops from a SoC if they wanted for the Mac Pro. It's actually better to have a SoC than to have a discrete GPU due to shared memory.
Because Intel has done a very good job at defining what a “Personal Computer” is over the past 40 years to the pony that most people really cannot see past Intel’s definition. Apple is going in a completely different direction, i.e. not a random bunch of parts put together on hope and a prayer, for which a lot of people Apple’s way is anathema.
There is also a class of users who will absolutely never be satisfied with the performance level of their computers. I call them Spec Chasers. They simply chase the next standard or spec or part and think that that is what defines a fast computer. This needs to a never ending churn of shifting allegiances to whoever has the fastest part at any moment in time. The average computer user just doesn’t care about all that. Hence why the MacBook Air sells like hotcakes. It’s not the fastest or the best at much of anything, but it’s overall a nice piece of kit and a great overall package for the average user. This forum is filled with a lot of people looking for wish fulfillment - from Apple, Intel, NVIDIA, AMD, Samsung, et al. basically whoever whets their appetite for the next big thing. It’s exhausting; they’re exhausting. A computer is a tool, just like a saw or a hammer. I don’t typically break up concrete with a framing hammer, but there are a lot of people here doing that. Or conversely, trying to nail a board together with a sledgehammer. Meanwhile, they buy the next shiny sledgehammer they see, and don’t get anymore boards nailed than the guy with the framing hammer. But, HEY, they got THE latest, greatest sledgehammer. Until next year.
Silly rabbit, it should be faster than an RTX 3090, because until it is, there’s a lot of people here for whom a GPU represents their manhood.
Last edited:
But those will never be satisfied with Apple and shouldn't spend their energy lamenting about Apple. 🙂
Some of the most expensive Apple computers were all headless Macs and displays in years before. The iMac all-in-one was a lot more economical then a Mac Pro with its built in display, that got better over time. We are now at that point in time that the next larger iMac will be pivoting from intel to M1 Pro/Max with a display quality uptick. The headless Macs currently are at extreme ends of the buying spectrum. Either the lowest cost, or the highest cost right now, if you don't look at clamshell MBP with displays being utilized.
Here's where I'm coming from:
Broadly, I'd divide Apple's Mac users into three classes:
I. Light/casual: Email, web browsing, basic office work. The M1 is designed for them.
II. "Power users":Large, complex documents in Word or LaTex; large, complex spreadsheets; math/science/engineering programs like Matlab, Mathematica, Maple, SaS, R, etc.*; protyping work for simulations, etc., that will be sent to a computer cluster. Their workflows don't require lots of cores. Instead, they require an intermediate no. of cores (say 8-10), plus high per-core speeds. As you'd expect, many who have STEM professions find themselves in this category; but I'd imagine there are many other classes of work that have these needs. [*A limited number of operations in Mathematica can be multi-threaded, but not most; I expect that's the case with these other programs as well—or it's possible you may be able to completely multi-thread some of them, but don't want to take the time to configure your program for multi-threading early in the prototyping stage.]
III. "Professional creatives and other heavy multi-core users": Those who have workstation-class tasks that require high core counts, and which they either can't, or don't wish to, send to a computer cluster, and thus instead do locally. Or those whose prototyping requires testing their programs locally with many cores. The Mac Pro is designed for them.
Now let's consider the broad category of users in Class II. Most of the programs they use are single-threaded. When they buy a desktop, the'll typically buy an i9 iMac (or maybe an i7 if they have limited funding). Yet even with an i9, they can often find themselves waiting for the computer to respond. There are two classes of wait times:
A. Short wait times for repeated tasks. Sometimes the wait time can be only a couple of seconds, yet a couple of seconds for a repeated task can be irritating. They'd really like the pleasure of a computer that responds so fast they don't notice a delay. To achieve that, they'd need single-core speeds at least an order of magnitude higher than what the i9 provides. [Obviously this isn't just the CPU, but also everything else that supports it.] This gives you a picture of the enormous way we have to go in improving user experience when it comes to single-core performance. I.e., computers will have to get far faster than they are now before user experience is not significantly impacted by single-core speeds. This is why I disagree strongly with the idea that "the M1 single core speed is enough".
B. Long wait times for individual tasks. An example of this would be a single-threaded calcuation or program that takes, say, 10 minutes or 100 minutes on an i9. Here there's no need for an involved explanation for why improved single-core speeds would be greatly welcomed.
Last edited: