We can believe ARM MAC possible because the ARM architecture performance has progressed well, chip design/fab improved lot and in last few years Apple built excellent chip design team to make it happen. Also ,with own 5G networking chip design team, future MACs will have 5G enabled on them. It is about competing(stay ahead) in market place with continued improving cost,performance and schedule.
As far as Software porting, Apple have done it before and now with all the resources on hand, it can be easier and faster than before.
Many times I said.. This so called ARM Mac is not for professional customer.
For the professional customer, you will get Intel or AMD CPU Mac.
This is only for the regular consumer level. Who does not need professional application which heavily rely on x86 code base.
Most on consumers can use iOS apps to do their daily working or entertainment purpose.
That is the reason, Apple drop 12' Macbook. Too expensive, not much better performance then iPad Pro.
I wonder if Apple introduce iMac with A series CPU then they will solder RAM & SSD as well or not. I think they will.
Also I wonder order to handle 4K or 5K panel, maybe A series CPU will get AMD GPU like Intel's hades canyon?
So what you are saying is the X86 ‘code base’ will need to exist forever? Just like the headphone jack and SCSI? Microsoft, Adobe, and the other ‘professional’ developers are totally incapable of coding for a new platform? Really? I wonder why Office and Adobe CS exist for iOS then.
If Apple does this, I wonder if they'll license the x86 instruction set from Intel? That would make the most sense for at least the first few years until devs can update software. It will take YEARS for this to play out if they did. Of course, they could use a "Fat Binary" like they did during the PPC-x86 transition... That way, the app bundle will be able to support both platforms. My question would be about performance. No ARM chip will be able to outperform the Xenon chips in the Mac Pro and iMac Pro. What about those systems? Will developers have to keep using Fat Binaries forever because of that? There are just so many questions, that I don't know if it makes sense or not.
Does AMD license its instruction set from Intel? Why would Apple do so?
If history is any guide, the very first generation of ARM Macs might not be the one to buy.
How so? Technology always gets better, so the longer you wait to buy a new computer the better your new computer will be. But aside from that issue, I don't recall first generations of new processor Macs from Apple as being somehow problematic.
I'm not thinking of issues with new processor families, but with new Mac product lines in general. Two recent examples: the 2016+ MBP (keyboards, 16GB RAM limitation); the first generation new retina MBA (underpowered; keyboards). There's been an Apple problem of first releasing "maybe good enough" vs. "great", and those are just two illustrations. And note that 4 years on the MBP still isn't fixed, although the recently released MBA is great...now.
As a longtime iPhone and iPad user (currently XS, and 10.5" iPad Pro), as well as the 2016 MBP I'm typing this on, I'm actually looking forward to an ARM transition. But I'm going to take a long, hard look at whatever it turns out to be before buying. (BTW, I buy several new Macs a year for work.)
We can believe ARM MAC possible because the ARM architecture performance has progressed well, chip design/fab improved lot and in last few years Apple built excellent chip design team to make it happen. Also ,with own 5G networking chip design team, future MACs will have 5G enabled on them. It is about competing(stay ahead) in market place with continued improving cost,performance and schedule.
As far as Software porting, Apple have done it before and now with all the resources on hand, it can be easier and faster than before.
So, if MacBooks are running on the same processors as iPads, is there then any difference?
Should MacBooks run iPad OS?
Or
Should iPads run MacOS?
Or
Maybe both rather than either/or?
In any case, the distinctions between two are becoming increasingly small. It may be why Apple added the USB-C port and finally acquiesced to reality and added cursor support to the iPad -- to start the transition to homogeneity.
I might hold off on buying a 2020 Mac mini and go for a discounted 2018. The lack of a CPU upgrade this year may mean there's an ARM mini coming out next year!
If Apple does this, I wonder if they'll license the x86 instruction set from Intel? That would make the most sense for at least the first few years until devs can update software. It will take YEARS for this to play out if they did. Of course, they could use a "Fat Binary" like they did during the PPC-x86 transition... That way, the app bundle will be able to support both platforms. My question would be about performance. No ARM chip will be able to outperform the Xenon chips in the Mac Pro and iMac Pro. What about those systems? Will developers have to keep using Fat Binaries forever because of that? There are just so many questions, that I don't know if it makes sense or not.
I don't think there is any chance Intel will license the x86 instruction set to Apple. I also think it's premature to declare that no ARM chip will be able to outperform Xenon. By avoiding Intel's 40-60% margins, Apple can throw 1.6-2.5x the silicon at the problem and maintain the same selling price.
These processors are surprisingly competitive with high-end Xeon and Epyc CPUs from Intel and AMD, yet they merely use standard core designs from ARM --- the kinds of standard core designs that Apple's A-series chips routinely obliterate in benchmarks, both synthetic and real-world.
Combine these data points and I think Apple definitely could make a CPU for the Mac Pro that blows away the x86 competition.
The big question is -- will they do it?
An argument against is that the Mac Pro is too small of a market to warrant the investment.
A stronger argument in favor is that Apple might have more uses for high-end ARM CPUs than just the Mac Pro. For example, suppose Apple were to go with a "chiplet" design as AMD has. AMD uses the same 8-core chiplet in a wide range of products. They just vary the number of chiplets that they package together. The high-end Threadripper 3990X that costs $4k has eight 8-core chiplets for a total of 64 cores. Meanwhile, the Ryzen 3700X has a single 8 core chiplet and costs about $280. So AMD can serve a wide range of needs using the same chip, the same design. Apple could do the same.
But even if Apple didn't use a chiplet approach, the market for a high-end chip could still be bigger than the Mac Pro because it might include Apple's own enormous data centers. I doubt Apple enjoys paying Intel for all the Xeon CPUs in Apple's data centers. Perhaps the largest customer for an ARM-based rack-mountable Mac Pro will be Apple itself.
“Apple to debut multiple ARM-powered Mac desktop and laptop models in 2021”.
This is written as fact when it’s actually just more speculation and clickbait. I don’t care if his track record is 100% accurate, until Apple announces it, it’s pure speculation and vapor ware. I expect better from the editors of Appleinsider.
All Apple tech blogs are rumor mills. Very little hard facts are presented. Rumors, editorials, opinion pieces, reviews, that’s about it. That some people reading those rumors take them as fact is sad. That happens a lot here. On the other hand there’s nothing wrong about daydreaming. I’m still waiting for the real Apple TV (big screen 4K TV) that Kuo has been predicting around the corner for years now.
There are lots of hard facts in rumours. Wait a second… I was thinking of hardly facts.
These processors are surprisingly competitive with high-end Xeon and Epyc CPUs from Intel and AMD, yet they merely use standard core designs from ARM --- the kinds of standard core designs that Apple's A-series chips routinely obliterate in benchmarks, both synthetic and real-world.
Combine these data points and I think Apple definitely could make a CPU for the Mac Pro that blows away the x86 competition.
The big question is -- will they do it?
An argument against is that the Mac Pro is too small of a market to warrant the investment.
A stronger argument in favor is that Apple might have more uses for high-end ARM CPUs than just the Mac Pro. For example, suppose Apple were to go with a "chiplet" design as AMD has. AMD uses the same 8-core chiplet in a wide range of products. They just vary the number of chiplets that they package together. The high-end Threadripper 3990X that costs $4k has eight 8-core chiplets for a total of 64 cores. Meanwhile, the Ryzen 3700X has a single 8 core chiplet and costs about $280. So AMD can serve a wide range of needs using the same chip, the same design. Apple could do the same.
But even if Apple didn't use a chiplet approach, the market for a high-end chip could still be bigger than the Mac Pro because it might include Apple's own enormous data centers. I doubt Apple enjoys paying Intel for all the Xeon CPUs in Apple's data centers. Perhaps the largest customer for an ARM-based rack-mountable Mac Pro will be Apple itself.
These processors are surprisingly competitive with high-end Xeon and Epyc CPUs from Intel and AMD, yet they merely use standard core designs from ARM --- the kinds of standard core designs that Apple's A-series chips routinely obliterate in benchmarks, both synthetic and real-world.
Combine these data points and I think Apple definitely could make a CPU for the Mac Pro that blows away the x86 competition.
The big question is -- will they do it?
An argument against is that the Mac Pro is too small of a market to warrant the investment.
A stronger argument in favor is that Apple might have more uses for high-end ARM CPUs than just the Mac Pro. For example, suppose Apple were to go with a "chiplet" design as AMD has. AMD uses the same 8-core chiplet in a wide range of products. They just vary the number of chiplets that they package together. The high-end Threadripper 3990X that costs $4k has eight 8-core chiplets for a total of 64 cores. Meanwhile, the Ryzen 3700X has a single 8 core chiplet and costs about $280. So AMD can serve a wide range of needs using the same chip, the same design. Apple could do the same.
But even if Apple didn't use a chiplet approach, the market for a high-end chip could still be bigger than the Mac Pro because it might include Apple's own enormous data centers. I doubt Apple enjoys paying Intel for all the Xeon CPUs in Apple's data centers. Perhaps the largest customer for an ARM-based rack-mountable Mac Pro will be Apple itself.
Many times I said.. This so called ARM Mac is not for professional customer.
For the professional customer, you will get Intel or AMD CPU Mac.
Professional what? Over here in professional developer-land, we'll have the easiest time of anyone switching over. All this *nix stuff just needs a fresh make and you're off to the races.
These processors are surprisingly competitive with high-end Xeon and Epyc CPUs from Intel and AMD, yet they merely use standard core designs from ARM --- the kinds of standard core designs that Apple's A-series chips routinely obliterate in benchmarks, both synthetic and real-world.
Combine these data points and I think Apple definitely could make a CPU for the Mac Pro that blows away the x86 competition.
The big question is -- will they do it?
An argument against is that the Mac Pro is too small of a market to warrant the investment.
A stronger argument in favor is that Apple might have more uses for high-end ARM CPUs than just the Mac Pro. For example, suppose Apple were to go with a "chiplet" design as AMD has. AMD uses the same 8-core chiplet in a wide range of products. They just vary the number of chiplets that they package together. The high-end Threadripper 3990X that costs $4k has eight 8-core chiplets for a total of 64 cores. Meanwhile, the Ryzen 3700X has a single 8 core chiplet and costs about $280. So AMD can serve a wide range of needs using the same chip, the same design. Apple could do the same.
But even if Apple didn't use a chiplet approach, the market for a high-end chip could still be bigger than the Mac Pro because it might include Apple's own enormous data centers. I doubt Apple enjoys paying Intel for all the Xeon CPUs in Apple's data centers. Perhaps the largest customer for an ARM-based rack-mountable Mac Pro will be Apple itself.
Many times I said.. This so called ARM Mac is not for professional customer.
Professionals do utilize devices with AArch64 and they will continue to do so once they care in a Mac.
This is only for the regular consumer level. Who does not need professional application which heavily rely on x86 code base. Most on consumers can use iOS apps to do their daily working or entertainment purpose.
This seems to imply that an ARM-based Mac will only be able to run iOS apps, not macOS apps. That's absolutely not the case as that would not be a Mac at all.
Apple A-series are SoC's. not really needed in a desktop system... More than likely it'll be a stand-alone CPU dubbed the X-series, and of course G-series for GPU.
If Apple does this, I wonder if they'll license the x86 instruction set from Intel? That would make the most sense for at least the first few years until devs can update software. It will take YEARS for this to play out if they did. Of course, they could use a "Fat Binary" like they did during the PPC-x86 transition... That way, the app bundle will be able to support both platforms. My question would be about performance. No ARM chip will be able to outperform the Xenon chips in the Mac Pro and iMac Pro. What about those systems? Will developers have to keep using Fat Binaries forever because of that? There are just so many questions, that I don't know if it makes sense or not.
Does AMD license its instruction set from Intel? Why would Apple do so?
Yes, they cross-license from each other. AMD licenses the x86 ISA and Intel licenses the amd64 ISA to give us what is now known as x64.
Comments
As a longtime iPhone and iPad user (currently XS, and 10.5" iPad Pro), as well as the 2016 MBP I'm typing this on, I'm actually looking forward to an ARM transition. But I'm going to take a long, hard look at whatever it turns out to be before buying. (BTW, I buy several new Macs a year for work.)
Should MacBooks run iPad OS?
Biz-Book
Personally I think Apple will release:”the business book” with ARM first to replace the outdated small MacBook.https://www.anandtech.com/show/15621/marvell-announces-thunderx3-96-cores-384-thread-3rd-gen-arm-server-processor
https://www.anandtech.com/show/15578/cloud-clash-amazon-graviton2-arm-against-intel-and-amd
These processors are surprisingly competitive with high-end Xeon and Epyc CPUs from Intel and AMD, yet they merely use standard core designs from ARM --- the kinds of standard core designs that Apple's A-series chips routinely obliterate in benchmarks, both synthetic and real-world.
Combine these data points and I think Apple definitely could make a CPU for the Mac Pro that blows away the x86 competition.
The big question is -- will they do it?
An argument against is that the Mac Pro is too small of a market to warrant the investment.
A stronger argument in favor is that Apple might have more uses for high-end ARM CPUs than just the Mac Pro. For example, suppose Apple were to go with a "chiplet" design as AMD has. AMD uses the same 8-core chiplet in a wide range of products. They just vary the number of chiplets that they package together. The high-end Threadripper 3990X that costs $4k has eight 8-core chiplets for a total of 64 cores. Meanwhile, the Ryzen 3700X has a single 8 core chiplet and costs about $280. So AMD can serve a wide range of needs using the same chip, the same design. Apple could do the same.
But even if Apple didn't use a chiplet approach, the market for a high-end chip could still be bigger than the Mac Pro because it might include Apple's own enormous data centers. I doubt Apple enjoys paying Intel for all the Xeon CPUs in Apple's data centers. Perhaps the largest customer for an ARM-based rack-mountable Mac Pro will be Apple itself.
https://appleinsider.com/articles/20/03/26/the-groundwork-is-set-for-apple-pro-arm-mac-chips
Thanks for the link!
An iPad Pro can already run it's own screen plus a 4k display over USB-C, so Apple GPU already run the equivalent of 5k panels.
This seems to imply that an ARM-based Mac will only be able to run iOS apps, not macOS apps. That's absolutely not the case as that would not be a Mac at all.
Apple A-series are SoC's. not really needed in a desktop system... More than likely it'll be a stand-alone CPU dubbed the X-series, and of course G-series for GPU.