Me too , in the meanwhile to save investenets maybe our next setup will be , Ipad Pro , Mac mini Arm , Intel Nuc.
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.
A14X Chip for First Apple Silicon Mac and New iPad Pro to Enter Mass Production in Fourth Quarter
- Thread starter MacRumors
- Start date
- Sort by reaction score
Anything less than 100 cores will be too laggy. Even for an entry-level machine. Need a minimum of 645 GB of RAM too, or you won’t be able to browse the Internet properly.
Last edited:
There are major engineering challenges to multi chip that make it not worth their effort. I designed opteron which was designed to work that way. It was a necessity because we couldn’t put more cores on a die, but it is a pain and not worth it given the external bus speeds.
Why are loads of people on this thread using extremely high end computers to compare the first AS Macs to & saying if the first chips out of the door, don’t turn Apple has failed.
Apple is obviously going to first target the machines that they sell the most of - which by all accounts are the MBA (Overwhelmingly so), 13 inch MBP & 21.5 inch iMac.
Conveniently for them, the machine that sells the most, has the lowest end processors, which is then a sensible place to start.
All they have to do then, is to beat the i3 & i5s in the MBA, make the integrated graphics better and the battery life better (and make it fanless).
By all accounts, they will easily achieve this.
When we see the iMac and the 13-14 inch MBP things will get interesting as we’ll get an idea of how things will scale.
Apple is obviously going to first target the machines that they sell the most of - which by all accounts are the MBA (Overwhelmingly so), 13 inch MBP & 21.5 inch iMac.
Conveniently for them, the machine that sells the most, has the lowest end processors, which is then a sensible place to start.
All they have to do then, is to beat the i3 & i5s in the MBA, make the integrated graphics better and the battery life better (and make it fanless).
By all accounts, they will easily achieve this.
When we see the iMac and the 13-14 inch MBP things will get interesting as we’ll get an idea of how things will scale.
It may be preferable in a power is not an issue scenario, but you don’t want faster cores that sacrifice battery for example. That’s why we’re seeing more emphasis on hybrid cores now.
Also I think you are hyper focused on single threaded applications and not taking a system look. More cores and threads means less bottle necking and context switching. Single cores can be allocated to different tasks, same for threads. In a age where a lot of pulling is happening for notifications for example it’s advantageous to have multiple cores.
In addition to that- a lot of other tasks like tabs in Chrome and Safari use their own threads. Combine this with video decoding it’s clear a multi core advantage.
We’ve been enjoying increasing multi core performance for awhile
I feel like people are over hyped and maybe hoping Apple breaks out of Moore’s Law or something. But by god physics will prevail xD
this is a great conversation but I think we are forgetting the marketing angle of it - it would make no sense for Apple to come out guns blazing and ship its first Mac with 16-20 cores with 3-4Ghz. then what would they sell for next year's models. IMO, I expect Apple to handly beat anything from Intel/AMD with less power consumption and longer battery life but the features/performance will be dribbled out into multiple SOC generations.
Last edited:
Apple won't fail if they start off on the low end, I can't imagine anyone is thinking that as we know CPU in the dev kit is performing very well. What will be disappointing is if the Apple CPU is only in the low end from day one as it will continue to push this perception that ARM is only for the low end machines. A halo machine with the best of what Apple can do with ARM(That is more than just matching the current crop of Intel CPU's) is what I am sure a lot of people are looking forward to seeing. We already had Lightroom and Photoshop demo'ed... they aren't for a MacBook Air.
then you will be thrilled since the same application and services will likely run across the watch/phone/Ipad/MAC
If next year they don't have even faster machines, isn't that not just as bad as sticking with Intel? I really want to see them push out something where they can say 2x in Photoshop, 3x in Final Cut Pro X 8k timelines, even if that is 'cheating' by using co-processors.
I get that 1:1 performance with more battery life and less heat is a win, I really want to see Apple push something next gen. We have been stagnant in computing for nearly a decade. My 2010 MacPro should not even have the grunt left in it to try to open a game released today.
I can understand Digitimes - they follow the supply the chain -they would hear about Apple is buying xxxx parts but where they go wrong is that they will extrapolate this to then to say Apple is shipping xxxxx and they are almost always wrong
On Jon Prosser - his sources got the date wrong but the information was ultimately correct - the ipad and watch will shipped first independently from the phone which a lot of people said it was 'impossible' because they never done that before.
Last edited:
ARM core architecture will arm Apple with gazillion ways to release a new Mac every quarter. Just add few more cores and safari will be snappier on every release.
Apple is going to get there with its processors but it is not going to be overnight. Apple is starting right with the low end of the MacBook models and moving up the line as they can improve CPU processing and GPU processing. Henry Ford did not build a 8 cylinder turbocharged OHC engine with 1000HP for his first car, he started with a 4 cylinder engine and he still sold thousands. Is Intel going away as a processor company no, AMD going away as a processor manufacture no. You ramp up as you work out the details.
I have my 2019 MacBook Pro 16 with 2TB of storage and 32gb of ram I got last fall, it is going to last me a while as Apple works out the details of it's new system. I will get a new Mac with a new Apple processor but it is not going to be next year and maybe the year after.
I mean, sure, but the assertion was "Gonna need a lot more than 8 performance cores for it to be useful". Highly doubt you'd accomplish that without sacrificing battery life.
(Plus, they're talking about desktops.)
Right.
Yes, that's true up to a point, but you're not gonna saturate more than a handful of cores that way, even today. I don't think a 8+4-core MacBook Air, as Gurman suggests, will saturate those cores much at all.
Android can make use of more cores by putting the garbage collector and JIT in those threads, so that's different.
We’ve been enjoying increasing multi core performance for awhile
Multi, yes.
24, as that poster was saying? Not for almost anyone.
Oh, nice!
Does Trip Mode set that? I assume Instant Hotspot sets it automatically?
yes- Apple SOC's will be faster but not in the way we think traditionally - it won't be like Machine A runs AMD/Intel with xx cores so then you would assume this will beat an Apple SOC. Apple has workflow team who has the inside of what is needed to process video/audio editing, photo/storage management, etc. and they will build accelerators for that. I expect the fanless 12" to do this job without a hitch while Machine A will struggle, generate heat and blow the fans
No. The A14X is entering production in Q4. The A14 is already in production (A14 goes into iPhones) . It is one reason why the A14X will be limited to less than 6000 wafers because the A14 is soaking up most of the wafer starts. There is no "iphone powered Mac".
Additionally, the A 14x is a bigger chip (minimally a bigger GPU (more GPU cores) and system memory/cache. if not also more CPU cores ) . Which means you get less chips per wafer processed. ( It needs more wafers to do very high volume. So that is tough when competing against the phone which needs an order of magnitude more chips. )
It is going to be touch for Apple to launch new iPhones and new Macs at the same time which all use the latest, bleeding edge process. ( adding in iPad Pro doing the same thing will be even harder. )
Unless Apple is reviving the USB-C only (not Thunderbolt) , 1-2 port Macbook , there may not be a sharing of the A14X between iPad Pro and MacBook. This A14X rumor could be just the allocation of rumors for an iPad Pro ( which might start shipping in 2021 depending upon how late this Q4 start is. )
Basically it'll be an 12-core or 4-core i3 depending on the power mode. I'll pick one up to replace my aging mbp, as long as it has 16gb of ram or more.
AMD's 4800/4900HS have been in the market for a while now (January 2020 IIRC). Until the Tiger Lake announcement AMD was about as fast single-core as Intel's best laptop CPU offerings so to say they had caught up is not entirely wrong.
This has changed now, of course, but surely AMD hasn't been standing still all those months. And while Tiger Lake CPUs are faster single-core (according to Intel) they have only released dual core and quad core chips so... probably not going to be great in multi core.
Two major problems.
First, macOS isn't all that great at NUMA. Coupled to that Apple has really about zero interest in forking off the macOS kernel to wider separation away from the iOS and iPad OS shared foundation. So the era of "multiple CPU packages" for Macs is very probably over. Done. Finished.
Second. The A-series chips are "sysyem on a chip". In part that there is now also "half a kitchen sink full" of other stiff on the die also. That means there is far more than just 'CPU' cores on the chip. What is the Mac Pro going to do with 4 different security enclaves? Four different "Unified memory system caches" ( how "unified" is it really if spread out over 4 different dies? ) . Four different independent GPUs. Four different memory controllers.
How do they hook those together ? ( that pragmatically involves more pins (whatever scheme come up with) which makes the chip physically bigger. Size matters. The iPad Pro is going to drive real limits on how big the SoC can be.
[ Hooking up multiple CPU packages gets worked out for server CPU products, but they are also physically much larger than mobile, tablet chips. It is doable, but it is also not "free" (trade-offs to be made. ) ]
The I/O bandwidth needs of the Mac Pro ( 6-8 slots and over half dozens ports ) is also vastly different than the two port wonder iPad Pro. That isn't going to scale across the same reused chip package either.
So Apple covering the whole (relatively much broader Mac line up versus iPhone or iPad) with just one chip. Quite unlikely. Apple will probably need multiple dies that fit into at least as wide spectrum of chip packages. The "one size fits all " and "hand me down" modus operandi of the iPhone and iPad products isn't going to fly across the whole Mac line up.
Thing is we are moving to the 14th generation now so we should have some expectation that Apple have a desktop class chip ready. And I really hope they don't just go low end first, that will do nothing to sway the publics opinion that ARM are for low cost and low end machines. We even see it in this forum where the expectation is that this will be a iPad with a built in keyboard and macOS(that now looks suspiciously more iPad like).
I really want to see them push a cracker of a product. A new iMac line would certainly do that more than a MacBook Air.