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.
Does anyone think there will be a quick 2nd gen of the M1 Macs that were just released?
- Thread starter TheArgonaut
- Start date
- Sort by reaction score
The MacBook Air, 2-port MacBook Pro, and Silver Mac Mini will not likely get updated until the end of 2021 or early 2022.
What we are likely to see sometime this year is an update to the 4-port 13” MacBook Pro, the 16” MacBook Pro, the iMac and space grey Mac mini.
Apple said at WWDC last June that the full transition will take about two years, so they have roughly until June 2022 to transition their entire lineup. It’s likely that the final Mac to fully transition over will be the Mac Pro, and their other higher-end desktops.
What we are likely to see sometime this year is an update to the 4-port 13” MacBook Pro, the 16” MacBook Pro, the iMac and space grey Mac mini.
Apple said at WWDC last June that the full transition will take about two years, so they have roughly until June 2022 to transition their entire lineup. It’s likely that the final Mac to fully transition over will be the Mac Pro, and their other higher-end desktops.
Personally, I decided that I’m waiting for more GPU performance from ASi Macs still.
What they’ve done is pretty great, but the dGPU on my Hack still eats it alive in that one regard.
A Mac Mini Pro (or iMac maybe) with more GPU muscle would finally get me over at some point I’m pretty sure.
What they’ve done is pretty great, but the dGPU on my Hack still eats it alive in that one regard.
A Mac Mini Pro (or iMac maybe) with more GPU muscle would finally get me over at some point I’m pretty sure.
Correction: it's a "low-end" SoC that competes head to head with Intel-based Macs (even the 16" MBP) in many applications. There's a reason MaxTech posted a video outlining why he replaced his $15,000 Mac Pro (12-core Xeon, 192GB RAM, Vega Pro videocard) with a 16GB Mac Mini. You're correct that up coming variants of the M-series will be even more interesting, but calling the M1 "low-end" is misleading when you consider how well it compares to current Intel CPUs.
To add my view, short answer is "No". They won't update the ones released last November until they have updated all of the other Macs to Apple Silicon.
For the M1 MacMini, I think this will take even longer as Apple will want people to buy an Apple Silicon version of the MacPro, not an upgraded Mini. I think there won't be an updated MacMini until a year or two after a new MacPro.
I'm still wondering if there will be a Mac Mini Pro in a black chassis again.
That could very well be their solution for the headless xMac crowd that wants something between the MM and MP, but that isn't attached to a screen (iMac).
There is definitely room in the MM chassis to do more if needed on cooling and/or larger or more hardware components.
That could very well be their solution for the headless xMac crowd that wants something between the MM and MP, but that isn't attached to a screen (iMac).
There is definitely room in the MM chassis to do more if needed on cooling and/or larger or more hardware components.
I think we might not see base MBP and MBA updated this year or if so not til end of year. Maybe they'll stay a Silicon gen behind, i.e. When the 3rd iteration of M-chips comes along, they'll get the 2nd.
The 4-port higher priced 13"/14"/16" will have M1x or M2 or whatever, as the upsell.
And iMacs will have it too, since they haven't been Silicon refreshed.
Mac mini, who knows. It'll get abandoned for a few years. Jk.. but I have no idea what cycle that will be on. It's an unpredictable product line.
The 4-port higher priced 13"/14"/16" will have M1x or M2 or whatever, as the upsell.
And iMacs will have it too, since they haven't been Silicon refreshed.
Mac mini, who knows. It'll get abandoned for a few years. Jk.. but I have no idea what cycle that will be on. It's an unpredictable product line.
Last edited:
Don't think there's going to be any drastic tick-tock changes to IPC or node shrink. 5nm is going to be around for probably another two years. If anything, maybe increase from 4 to 6 high power cores and new form factors like Surface detachable with touch and pen inputs.
I'd expect that the 2021 A15 and whatever Mac SoC is co-developed to use TSMC's 5nm+ which gives a moderate density increase over 5nm. After that, it will probably depend on how well TSMC's 3nm effort is progressing. Right now they are being extremely optimistic and saying full 3nm production in 2022. I'm skeptical but that would be great because that means the top end Macs will probably be based on the 3nm SoC/CPUs. I think TSMC is still talking about a 4nm node as well which might be the fallback node if 3nm doesn't make it by September 2022.
Like in days of yore, the mini will resume its role as parts bin queen. It will probably be the last M1 to get upgraded.
Again, more cores is not necessarily better performance. ARMs tend to show an edge in single core but trail in multicore benchmarks, so far. I could see them develop a hybrid core design that can do either big or little throughput, allowing the SoC to operate as all big under heavy loads or using little mode under light loads. But going past eight CPU cores is not linear scaling by any means – the cost/benefit ratio goes down pretty fast (though fattening the GPU is a different matter).
What is interesting is the neural engine, which packs a lot of potential. If Apple works on developing that functionality more (especially with more programmer-friendly APIs), the number of CPU cores might start to become even less significant if the neural thing can reel in the slack. Things like graphics and video editing will become all but effortless, with logic that can see things the way we do. This is what I consider the promising avenue for growth and enhancement in the M-series, and why schools should be interested in Macs because neural net program design may well overtake traditional programming sometime this decade.
What is interesting is the neural engine, which packs a lot of potential. If Apple works on developing that functionality more (especially with more programmer-friendly APIs), the number of CPU cores might start to become even less significant if the neural thing can reel in the slack. Things like graphics and video editing will become all but effortless, with logic that can see things the way we do. This is what I consider the promising avenue for growth and enhancement in the M-series, and why schools should be interested in Macs because neural net program design may well overtake traditional programming sometime this decade.
The performance benefit of adding cores can also depend on how wide the decoding goes. One of the biggest advantages of the M1 is that since the bytecode instructions are all the same size, Apple could go wide with the decoders. For x86 machines (where those instructions are variable in length), 4 decoders seems to be the upper limit without introducing additional complexity into the decoding operations of the processors. AMD is already on record as stating that 4 seems to be the upper limit. Since the M-series does not add complexity to this process with each additional decoder, it theoretically could go 16, even 32-wide. I believe that the M1 has 8 decoders (which would be 1 per CPU core) plus a huge reorder buffer to handle out of order execution. There is no technical limitation in place which would prevent Apple from going 12-wide on an 12-core M-series, or even 16-wide with a 16-core SoC.
I'm actually the same here, torn between diving in or waiting for the supposedly 14 inch Pros with MagSafe (which I have to say sounds sooo good). Got a late 2015 iMac and I'm currently about to start developing an iOS App which I'm almost certain this iMac won't be able to handle.
All I know is when I'm 50/50, I somehow always end up diving in ahaha.
Decoders themselves are very simple logic arrays on an ARM CPU. There are a few layers of complexity to the instruction word, but they are all quite straightforward. The more important part is the dispatcher, which has to determine whether the instruction has pending dependencies (usually whether the required registers are being modified, or whether an out-of-order modification to the CCR would affect another instruction). Dispatchers manage instruction flow and internal parallelism to maintain a reliable machine state for the CPU. How many of those there are might be more interesting than decoder width (and is it possible/practical to balance more than one dispatcher in a single thread pipe, and do you gain by doing so).
I had that thought and bought one anyway. My MacBook Pro 13 AS is faster than my MacBook Pro 16. Also I can play Mankind Divided better than with the 16 including with an eGPU (granted only an RX580.)
These are stellar machines! I recommend it for anyone. These will only get better and better over time.
When Apple decides to schedule another event? They don't exactly have a set release schedule that we can plan for months in advance.
I don’t care about performs in synthetic benchmark. It only has 2 usb-c ports which apparently do not even work at full speed in comparison to Intel. It also does not even support multi monitors natively. Bluetooth also does not perform like Intel chips. And it is only limited to 16gb of RAM, unlike Intel chips.
So really, there are too many compromises, so it is a low-end chip. Also the M1 chip, looks like something that could have been put in the iPad Pro. It’s basically a rebranded A14X.
The 16” MBP hopefully addresses all the shortcoming of the M1 and then it will be a high-end chip, which is what I will be getting if Apple does their job.
The MaxTech video I referred to wasn't comparing synthetic benchmarks, they were comparing actual workflows in FCP, Lightroom, DaVinci resolve, etc. In most of those non-synthetic, real-world tests, the M1 either performed on par with or outperformed their Mac Pro. The one exception was with some of the higher end Red formats, where the dedicated GPU in the Mac Pro did all of the heavy lifting instead of the CPU and system RAM.