Whatever happens, there must be more sustainable development with software and gaming companies, to make this ARM technology more compatible. It's amazing the battery autonomy of the ARM laptops in usage, a little less in sleep or hibernate, but this is truly the future of ultrabooks, we have reach the limits of thermal dissipation, not only in laptops, but desktops also, with the old school thinking of Intel, Nvidia and AMD, a new mentality must to expand and dominate the market, so that we can all go forward. These laptops also benefit from great thermals, to say the least, if it wasn't for the low compatibility with software and games, I would already have purchased one. If all goes well, in 2-3 years software, 3-5 years for games, everything should be much better.

Hahahahaha

iSheep vendor mentality ;D ;D

I think there are two key benefits for Apple in making the transition to ARM.
1) The biggest reason: The Mac will access to hundreds of thousands of applications built for iOS and iPadOS. This will inevitably increase the likelihood of developers designing applications for Mac, as they will also be compatible with iPads and iPhones. Lower development cost, higher user base and more opportunity.

2) Apple can control their own chipset design for better optimisation, without relying on Intel who have been slower to make improvements. Intel improvement (multi core) from 5th gen to 10th gen i7 (5+ years) is up 69% according to Userbenchmark, and Apple A8x to A12x processor multicore performance is up about 350% according to Geekbench (in 4 years).

Seems like a lot more potential in the ARM architecture with faster progress, which could surely flow into Windows based apps, and also benefit Android.

Quote from: xxx on July 14, 2020, 22:10:49
Sorry, but the distinction between CISC and RISC has been irrelevant for at least past 20 years.

Modern "CISC" CPUs like x86 have been long (specifically since Pentium Pro in 1995) using RISC-like microcode for the actual execution was while modern "RISC" CPUs like ARM have adopted complex CISC execution for e.g. vectorization.

So in their internal design both design traditions have converged to largely similar approach. ISA is of course different but that just does not have that big of an effect nowadays.

Yes, I know. Both approaches have their merit and real processors have been converging. I'm not that familiar with ARM, but I do know of, for example, NEON. That was my point - they're more alike than many people think which is why I wouldn't expect a miracle. But that's the fundamental difference and where they're coming from. And one place where you see a difference between them is the front-end which is tied to their legacy.

A more in-depth comparison would be interesting. As far as I know (and I'm no chip designer), x86 processors are suffering quite a bit from legacy. While Apple was able to cut away legacy in their ARM processors (and the legacy wasn't that heavy in the first place). How much difference does it make, I don't know. It's certainly costing them transistors and therefore power. Or how much difference their different approach to memory consistency makes. I would quite enjoy reading it. These articles on "ARM good, Intel bad" typically lack any insight and interesting information.

macbook is already much slower than windows laptops in everything, the ARM based mac will be even slower, lets see how fast is that ARM(ipad cpu) compare to amd 4900 or intel i9, they are nowhere... Lets hope windows not make any ARM based bs... Yes battery life... the only thing apple get is they dont need to pay intel for cpu(which is not cheap), what we can get is an ipad wich could run all the osx apps, which would be great though...

Quote from: Theo on July 14, 2020, 17:21:05
An apple veteran should very well know the upmost important factor of battery life is manufacturer / oem / electrical engineering.

Not the cpu or chipset or screen or even battery size.

Design wrong and you will waste 100 Wh in an hour on 2 core arm - and most oem's still can't engineer that on at least some of their devices.

Well, it's possible to screw up everything. Good components don't guarantee a good product. But they set the baseline. They limit how good it can be. And everything can last with a big enough battery. Making a battery bigger is the brute force approach to endurance. Another important factor is BIOS and drivers. There have been x86 laptops lasting around 15 hours in the Wi-Fi test (well, I recall at least one). Yes, OEM's competence plays major role.

I don't think it's reasonable to expect miracles. I'm open to being positively surprised. But I'm not holding my breath. Current x86 processors are far from trash. And Apple won't really give us answers. It wasn't easy to compare with Apple even on x86. Too many differences. And they're not going to sell their processors to competition anyway (A13 certainly shows promise looking at SPEC2006 results; Qualcomm, Samsung and co. are far behind). What I have seen in server space for now doesn't fill me with confidence. They tend to cherry pick bad competitors for their comparisons to make themselves look good and when you adjust for the poor choice, it's not very convincing.

Quote from: _MT_ on July 14, 2020, 20:23:12
Just because ARM processors you're familiar with have been designed for mobile devices doesn't mean that you can't design a powerful processor for a desktop. The principal difference between RISC (ARM) and CISC (x86) is that CISC offers instructions that can do complex work, taking up multiple cycles while RISC focuses on keeping instructions short and more complex tasks are broken up into smaller pieces.

Sorry, but the distinction between CISC and RISC has been irrelevant for at least past 20 years.

Modern "CISC" CPUs like x86 have been long (specifically since Pentium Pro in 1995) using RISC-like microcode for the actual execution was while modern "RISC" CPUs like ARM have adopted complex CISC execution for e.g. vectorization.

So in their internal design both design traditions have converged to largely similar approach. ISA is of course different but that just does not have that big of an effect nowadays.

Quote from: Superguy on July 14, 2020, 16:50:19
x86 isn't going anywhere. Apple has only a small slice of the PC market, and couldn't possibly meet demand. x86 chips have improved a lot, are more powerful, and are deeply rooted in all levels of computer architecture.  x86 is designed more for performance first, with power being more of a secondary consideration. ARM is designed for power efficiency at the cost of performance. ARM systems could throw more cores out to increase performance.
...
20 hours of battery power sounds great, but who actually needs that much?  And will they really achieve it with normal use?  They can't do it now with phones?  I think that's a big claim that has yet to be backed up.

Just because ARM processors you're familiar with have been designed for mobile devices doesn't mean that you can't design a powerful processor for a desktop. The principal difference between RISC (ARM) and CISC (x86) is that CISC offers instructions that can do complex work, taking up multiple cycles while RISC focuses on keeping instructions short and more complex tasks are broken up into smaller pieces. The complex instruction can be potentially implemented more efficiently in hardware. But the processor is going to be more complex and difficult to design. It also complicates features like out-of-order execution. In reality, AFAIK, processors we have today are hybrids of the two approaches. They have been borrowing from each other's cookbooks.

Yes, it's entirely possible to replace x86. Whether it's going to happen, who knows. This isn't a question of "which is better?" It's not uncommon for the lesser product to win. When it comes down to it, the strength of Intel processors is hardware emulation of x86. The weakness is that the emulation is always there. As long as we're running x86 software, it's going to be difficult to beat x86 processors. Once we have ARM native code, hardware emulation can become a burden. That doesn't mean there is nothing Intel (or AMD) can do.

It can last over 20 hours playing video. Which is primarily a question of video decoder efficiency.

An apple veteran should very well know the upmost important factor of battery life is manufacturer / oem / electrical engineering.

Not the cpu or chipset or screen or even battery size.

Design wrong and you will waste 100 Wh in an hour on 2 core arm - and most oem's still can't engineer that on at least some of their devices.

Not to mention that you need appropriate processors. That's actually the first step. Then OEMs can be interested and Microsoft can add support for them. Without a suitable processor, Microsoft can support all it wants and OEMs can wish all they want, but nothing will come of it. x86 world is much more than just ultrabooks.

These articles are also funny because people often forget that Intel's processors are internally RISC machines. With, essentially, hardware emulation of CISC. This arrangement gives them a free hand in evolution of the internal instruction set. There are no backward compatibility issues as it's internal. Downsides are that there is a complex decoder taking up space and you don't have the option to run native code. It's the software that has been keeping x86 alive.

Personally, I would expect servers to be easier to crack than consumer PCs (with Linux everywhere). Not much success there. I'm really curious to see what Apple does with their more powerful Macs. As I said before, if anyone can do it, Apple can. Because of their vertical integration. I don't think a typical PC OEM has big enough balls for this. It would probably take a company like Intel to push through the change. But Intel is hardly going to do it. Perhaps Google or Facebook for their server farms, if they went open hardware.

x86 isn't going anywhere. Apple has only a small slice of the PC market, and couldn't possibly meet demand. x86 chips have improved a lot, are more powerful, and are deeply rooted in all levels of computer architecture.  x86 is designed more for performance first, with power being more of a secondary consideration. ARM is designed for power efficiency at the cost of performance. ARM systems could throw more cores out to increase performance.

Apple has been more for specialized tasks like photo and and media editing where they've traditionally excelled.  They've also been more about form over functionality - they'd rather have a device that looks good and they're willing to sacrifice performance to get there. So it makes sense that they'd look toward ARM.  If someone is just looking for a PC to do web surfing, document editing and watch some videos, it doesn't make much sense to pay Apple's tax now, and I don't see prices coming down with a move to ARM architecture.  If anything, prices will increase as all their software needs to be ported over, and they'll be maintaining software for two architectures.  That isn't free.

20 hours of battery power sounds great, but who actually needs that much?  And will they really achieve it with normal use?  They can't do it now with phones?  I think that's a big claim that has yet to be backed up.

I'm not buying the argument for ARM taking over the world.

That guy forgets about one thing - x86 CPUs improved energy efficiency a lot in recent years. Now  there is no problem to find laptop that can run for 5-6 hours without outlet and that is enough for big majority of users.  Performance-wise ARM wont be able to compete with x86 for few next years. So yea, it may be future, but as things are now  I doubt I will consider ARM CPU when I will upgrade my current laptop.

It's a tall order, calling that no one will be able to compete with ARM based Macbook if Microsoft doesn't released an fully fledged Windows for ARM.

He forgets something way more bigger than Windows: Linux/GNU and Android, which is already running on ARM, and almost everywhere.