Macs Could Jump to ARM in 2020

Interesting. If so another kick in the pants for Intel. Sheeple don't care whats in the Macbook as long as its really expensive and fashionable... ;)
 
nutzo said:
This should really be helpful to all the people who need to run Windows apps on their Mac. :rolleyes:
(not)

This will largely make the Mac unusable for many business people, since any Windows emulation will be much slower on a non intel CPU.
Click to expand...

It depends, but tons of people I know would be perfectly fine if Adobe Creative Suite got ported over, along with things like MS Office (I'm sure it will, MS even has an ARM version of Windows).

A lot of software devs will be forced to buy the new Macs as well if they target iOS. I won't be too happy with the switch, right now I'm able to run macOS on a PC for my iOS dev needs, but in the future it looks like I might actually have to get a Mac laptop. I imagine they'll still support older devices for awhile though
 
Elios said:
software and backwards compatibility is the big issue emulation of x86 would be SUPER slow and useless for any thing other light aps thats the really the only draw back.
Click to expand...
How exactly did Apple do this, then, when they moved from m68k to PowerPC in the early 1990s, and then from PowerPC to x86 (a year later x86-64) in the mid 2000s??? (hint: universal libraries for both ISAs and/or emulation)
Programs can be recompiled and optimized for different ISAs, and it isn't like everything will be 100% locked into x86-64 instructions, especially if Apple starts giving the developers and software engineers SDKs and time before the move, which they obviously would.

ARM instructions are also emulated when running Android on x86-64 CPUs as well, and Microsoft also did this, to a much lesser extent, with Windows RT and x86/x86-64 backwards compatibility.
If Apple moves to ARM64 circa 2020, it will be far from game over, like everyone here who is touting "doom and gloom", is saying.

RISK has the potential to be much faster but your never going to run your old software on it at any real speed for a long time
Click to expand...
You mean RISC?
Never say never - emulation and dual-compatible software libraries have come a looong way since the 1990s and 2000s, and anyone who develops any level of serious or professional software for OS X is going to start recompiling, porting, and migrating their software from x86-64 to ARM64.

If they don't?
Well, the world of legacy software will welcome them, then. ;)
 
Last edited:
  • Like
Reactions: DocNo
like this
DNMock said:
So serious question here, if that were the case that an ARM chip could be scaled up to desktop levels performance, which I don't doubt that it can, there has got to be a major drawback to doing so other than on the software side otherwise there would be a push for AMD and Intel to switch to producing ARM processors for desktops, servers, etc. etc. and there doesn't seem to be one that I am aware of.
Click to expand...
That day is fast approaching quicker than nearly everyone here is realizing.
x86-64's days are numbered in the industry as a whole, and the only thing that has kept this nearly 40 year old ISA (x86) alive is backwards compatibility and the extreme (not so much these days) cost to move compatible software to a new CPU ISA.

x86-64 has nearly reached its limit, and will absolutely do so by 2030.
Cores and SMP software will only scale so far, the NM shrinkage can only get down to about 2nm (that is the physical limit of any classical CPU design, period), and software efficiency for x86-64 (especially compared to ARM64) is waning.

AMD is really the only company keeping x86-64 afloat in any serious capacity at this point, and Intel is literally limping along and trying to keep up while being competitive in a very mediocre way at this point.
Intel's constant hardware security blunders and the (task/scenario-specific) performance hits from it are not helping their case, either - Meltdown, Spectre, and Foreshadow all have serious performance hits when fully patched in certain scenarios, and especially in server environments like HPC and data centers.

The writing is on the wall for x86-64, and ARM64 is going to be the next major contender in the foreseeable future - this isn't an "if", it's a "when", and somewhere around ~2020 marks the spot.
Heck, ARM64 is even starting to make serious performance strides in server environments - just check out juanrga's thread on it.
 
Krenum said:
Faster than a i5 notebook processor....that's not saying much.
Click to expand...
It is when that iPhone 8 A11 ARM64 CPU has a 5 watt TDP - compared to that x86-64 i5 "notebook" CPU with a 28 watt TDP.
Yes, that is extremely telling.

The CPU in the iPhone 8 is also 40% faster, overall, than the 8-core AMD Jaguar in the PS4 Pro, which has around a 50-55 watt TDP (not counting the GPU).
I know the Jaguar is already over a half-decade old at this point, but again, that shows the strides that the ARM64 ISA has come in recent years.

Give that A11 SoC a 30-95 watt TDP - I would say that would make a pretty good contender against modern quad/six-core Core i5/i7 CPUs.
 
IdiotInCharge said:
Of a desktop implementation? Outside of a new Mini, it's doubtful we'd see such a thing for even longer. You still need desktop-level hardware for many tasks.

The overall point though is that below those tasks, you don't need a desktop CPU ;).
Click to expand...

After all, who cares if the tasks below that run slow, as long as they run, eh? :cautious::eek::rolleyes:
 
Krenum said:
Faster than a i5 notebook processor....that's not saying much.
Click to expand...

https://www.anandtech.com/show/13392/the-iphone-xs-xs-max-review-unveiling-the-silicon-secrets/4

"What is quite astonishing, is just how close Apple’s A11 and A12 are to current desktop CPUs. I haven’t had the opportunity to run things in a more comparable manner, but taking our server editor, Johan De Gelas’ recent figures from earlier this summer, we see that the A12 outperforms a moderately-clocked Skylake CPU in single-threaded performance. Of course there’s compiler considerations and various frequency concerns to take into account, but still we’re now talking about very small margins until Apple’s mobile SoCs outperform the fastest desktop CPUs in terms of ST performance"
 
This will probably happen. I'm sure Apple would love to move their ecosystem to a single architecture, and adopting x86 in phones isn't going to happen. They already have mature tools ready for developing on ARM, so the groundwork is already laid out.

And to the camp that doesn't think an ARM chip could ever compete with intel desktop silicon - well you're just plain wrong. It's just that no-one has actually tried to build one. Correct me if I'm wrong, but Intel's cores are just custom RISC hidden behind the monster x86 decoder. There is no technical reason that ARM cannot perform like x86.

In any case, we need competition in the desktop cpu arena - and unless hell freezes over we wont see another x86 contender, so its gotta be ARM.
 
H2R2P2 said:
You missed the underlying point I was getting at.. There is a reason Intel CPU's are the way they are. A lot of it has to do with maintaining backwards compatibility, but the other aspect is due to speed when being used during non-deterministic, purpose built purposes. For example, you can build a very simple RISC based calculator. It will be very cheap, very energy efficient, etc.. Now build a calculator using a traditional Intel CPU and corresponding computer parts you need to make it work. The computer based approach is comparatively extremely expensive, uses comparatively extreme amounts of power, isnt portable, etc.. Having said this, both will work perfectly fine for calculator duties; but if you ever need to do something with those numbers you are calculating (such as put them into a spreadsheet, database, share with another department, etc.) than the purpose built calculator is a joke.

ARM makes sense in the areas they are used. Yes, there are ARM based products (like the Chrome Book, Surface RT (I dont recall what the most current ARM based Surface is called) that provide traditional "computer" duties but they are woefully inadequate if you use them outside their narrow scope of intended uses.

If Apple goes towards ARM for their MacBook lineup, then they are throwing in the towel with regard to performance and will be a glorified Chrome Book. Heck, they switched *TO* Intel CPUs and away from Motorola (witch was also a RISC based design) because they weren't serious options for anyone other than in the sound/graphics/content creators.
Click to expand...

That is such a load... The A12 already runs circles around many Intel chips and does so with far less power consumption. At the rate Apple has been improving their processor designs, it was only a matter of time before this happened. Just imagine a few of these chips running in parallel. Just because Microsoft failed at ARM with their shitty development, doesn't mean Apple will.
 
  • Like
Reactions: DocNo
like this
IdiotInCharge said:
I'll provide a counter- Intel has, despite being stuck at 14nm, improved clockspeeds, reduced power usage, and increased core-counts for their mobile parts.
Click to expand...

It's still no where near the low TDP or performance of what Apple has with ARM. Not even close.

I love how the comments in this thread aren't disputing ARM isn't faster than Intel's mobile offerings, just that it's no big deal because it's "only" Intel's mobile offerings.

Apple hasn't reveled a desktop chip yet. People thought they were joking when they announced 64 bit for mobile but they did, it was real and others were scrambling to match.

There's no reason Apple couldn't release a desktop chip if they really wanted to. The real question is, do the benefits outweigh the disadvantages? Would it be worth the development effort? Right now it probably doesn't make sense on the desktop. In three to five years? If Intel continues at their current lackluster pace I think things could be in an entirely different place. I wouldn't put it past Apple to have something in the works. It may never see the light of day, but if it did why would that necessarily be a bad thing?

This kind of pressure is a good thing even if you never intend to ever buy a Mac. Look at how Intel is scrambling to keep up with AMD right now. It's kind of nice, no? Apple can always remind Intel to go talk to IBM about what happens if Apple's needs can't be met.
 
BloodyIron said:
Modern ms office for OSX is written for x86, it's no small task to re release it for ARM.
Click to expand...

They don't have to ARM dominates the world of computing now. its in every phone, every tablet, a handful of laptops. All they have to do is extend the functions of office made for iphones and this is surely going to be what apples plan is, don't make everyone redevelop x86 for arm just spur them to continue development for ios then they can release some mac laptop that runs on arm and talk about its magical ability to run ios apps. over a couple of year period they upgrade the whole mac line.
 
H2R2P2 said:
You missed the underlying point I was getting at.. There is a reason Intel CPU's are the way they are. A lot of it has to do with maintaining backwards compatibility, but the other aspect is due to speed when being used during non-deterministic, purpose built purposes. For example, you can build a very simple RISC based calculator. It will be very cheap, very energy efficient, etc.. Now build a calculator using a traditional Intel CPU and corresponding computer parts you need to make it work. The computer based approach is comparatively extremely expensive, uses comparatively extreme amounts of power, isnt portable, etc.. Having said this, both will work perfectly fine for calculator duties; but if you ever need to do something with those numbers you are calculating (such as put them into a spreadsheet, database, share with another department, etc.) than the purpose built calculator is a joke.

ARM makes sense in the areas they are used. Yes, there are ARM based products (like the Chrome Book, Surface RT (I dont recall what the most current ARM based Surface is called) that provide traditional "computer" duties but they are woefully inadequate if you use them outside their narrow scope of intended uses.

If Apple goes towards ARM for their MacBook lineup, then they are throwing in the towel with regard to performance and will be a glorified Chrome Book. Heck, they switched *TO* Intel CPUs and away from Motorola (witch was also a RISC based design) because they weren't serious options for anyone other than in the sound/graphics/content creators.
Click to expand...


A lot of people including you make a big deal out of the processor technology ARM, RISC CISC all this shit doesn't really mean anything now days processors are so complex and so large that the ideas of some simply idea of how to run something being inherently more efficient or better is garbage. All these guys implement and copy the features and advantages of the other guys and if any instruction or architecture is shown to be better the next guy will just copy it. An ARM processor in a modern smart phone is not some specialized calculator its an all encompassing system on a chip that does tons of different things on par with a modern desktop. The only difference is that they are smaller and focused more on being cheap once companies start scaling them up and they already are they are going to get competitive with x86. Its only a matter of time when they have so much going for them in volume already. Except now all the apps coded for phones will run natively on some new desktops and laptops and for a time that will be enough to bridge the software gap until those companies can get their apps mouse and keyboard friendly. And lets not kid ourselves a mouse and keyboard could work perfectly fine in a phone GUI with zero modifications. Literally click instead of touch your finger. Finally even though for some time x86 will have a raw performance lead there will be a massive market for people who don't care about that which is the massive majority of consumers to grow into until they get the power acceptable enough to replace their entire line. Even Microsoft is flirting with ARM again this year cause they know damn well this is coming sooner or later.

The only play intel really had was to make atom work in phones forwhatever reason they did not focus enough money and time on that and they have lost that battle.

To me the irony of this all is that apple was the last company that needed to go arm but they will likely be the first to actually do it. The company that should have been pushing the shit out of arm in laptops and desktops was google they should have when all in and got rid of chrome os and just made it android laptops.
 
IdiotInCharge said:
Do you even take yourself seriously?
Click to expand...
AMD is the only company truly innovating x86-64 at this point, and now Intel is trying to catch up, with which in terms of price/performance, they aren't doing; seriously, a 12% advantage core-for-core at a 160% price tag is not innovating.
Also, if those Apple x86-64 i5 notebook CPUs don't have 28 watt TDPs, then what do they have?

Got anything to actually counter my stance other than insults and jackassery? o_O
 
Last edited:
Red Falcon said:
AMD is the only company truly innovating x86-64 at this point, and now Intel is trying to catch up, with which in terms of price/performance, they aren't doing; seriously, a 12% advantage core-for-core at a 160% price tag is not innovating.
Also, if those Apple x86-64 i5 notebook CPUs don't have 28 watt TDPs, then what do they have?

Got anything to actually counter my stance other than insults and jackassery? o_O
Click to expand...

My God, you are being serious.

So:
Tell me about AVX512 on AMD CPUs.
Tell me about Intel's actual ultrabook CPUs.

And tell me how your AMD stock is doing :ROFLMAO:
 
IdiotInCharge said:
My God, you are being serious.

So:
Tell me about AVX512 on AMD CPUs.
Tell me about Intel's actual ultrabook CPUs.

And tell me how your AMD stock is doing :ROFLMAO:
Click to expand...
What does this have to do with anything in the thread?
Care to fill me??

I'm willing to admit I'm wrong, but can you actually provide useful info than just saying "are you serious?!"..?
That isn't helpful in the least.


EDIT:
I see that AMD's stock is down - ok, your point is...?
As for AVX512, outside of very specific HPC applications, I didn't think it was being used much by consumer-level applications and programs since the performance gains it makes are lost since the CPUs (from what I've read) have to drop their clock speeds, thus losing most of the performance gains from it; it also isn't offered in lower-tier CPUs like Intel's Pentium lines, so it really isn't a universal FPU function at this point. - again, your point is...?

As for Intel's ultrabook CPUs, yes, those are very powerful for their TDP, but again, if you up the TDP of Apple's in-house ARM64 CPUs, they could very well become a serious competitor in the future.
I think you misunderstood me, I wasn't saying ARM64 is flat-out beating x86-64 all-around at this point in time, but give it 2 to 5 more years, and we will be singing a very different tune at that point.


I would really appreciate it if you would fill me (all of us, really) in on what you are actually trying to get at rather than throwing out vague questions facetiously.
 
Last edited:
diehardman said:
only Skylake-X and Skylake-SP have AVX512 btw
Click to expand...
Which further begs the question of just what the point that IdiotInCharge is trying to make.
Hopefully he explains his reasoning and questions he touted in more detail, as I actually would legitimately like to know.

If it is just that "ARM64 CPUs aren't faster than all current x86-64 offerings", well, I would agree with him on that.
My point is that ARM64 CPUs and SoCs have a lot of potential, especially with what we are seeing at their current TDP and offerings when compared 1:1 to most (not all) x86-64 offerings, and will one day (much sooner than later) take over x86-64 entirely.

I would also think that most (again, not all) tasks that require AVX-512 would be better suited towards GPU compute for embarrassingly parallel tasks.
Maybe not, but it would be nice if IdiotInCharge would care to chime in and educate us all. :whistle:
 
Last edited:
  • Like
Reactions: DocNo
like this
Red Falcon said:
Hopefully he explains his reasoning and questions he touted in more detail, as I actually would legitimately like to know.
Click to expand...

That you're making poor comparisons. Especially when it comes to the outdated Intel hardware that Apple uses (for reasons inexplicable). And even if you compare Intel's current ultrabook offerings, we know they're behind on process- if you are going to bet on that situation to continue, you're betting on Intel folding. It's not a sound bet. Make honest comparisons, when I asked if you were being serious, that was honest. Comparing an Intel 28w CPU to an ARM SoC is wildly off base.

Red Falcon said:
My point is that ARM64 CPUs and SoCs have a lot of potential, especially with what we are seeing at their current TDP and offerings when compared 1:1 to most (not all) x86-64 offerings
Click to expand...

I've already made this point myself.

Red Falcon said:
and will one day (much sooner than later) take over x86-64 entirely.
Click to expand...

This I find especially silly; you can look at 'trajectories', but you're missing that ARM (and AMD for that matter) are catching up with the low-hanging fruit; further, your denigration of x86 versus ARM is also off-base. The two are more alike than they are different, and you could strip down x86 to be power-competitive with ARM the same as you could build up ARM to be performance competitive with x86. The barriers for both are ecosystem-based. Intel x86 CPUs have shipped in phones- how many Top 500 ARM supercomputers are there? It's a poor comparison, but it illustrates the point- and I do earnestly believe that there will be ARM supercomputers in the Top 500; I just don't believe that the reason will be the instruction set.

Red Falcon said:
As for AVX512
Click to expand...

Last one: you're talking about true innovation. AMD isn't innovating except with respect to themselves; they're working on catching up to Intel. And they're still catching up to four year old Intel tech. Good for AMD! But it's not 'real innovation'. AVX512 is one example; Xeon Phi is another, if we're talking just about 'x86 innovation' and not AMD vs. Intel, because AMD seriously loses the overall competition. Their current limelight is largely made possible by an expanding market that Intel cannot fully satisfy combined with a willingness to sacrifice margins for market share- something Intel is pretty stubborn about.

Red Falcon said:
I see that AMD's stock is down - ok, your point is...?
Click to expand...

Really last one: I don't give two shits about these companies' stocks. My point is that your apparent ignorance comes across as AMD worship to such a degree that it looks outright intentional. I'll give you the benefit of the doubt that that's not the case, but I'm telling you about it so you can see what you've written from that perspective.
 
And I want to throw this out there: I'm not anti-AMD. I'm skeptical, because I've run AMD, I've seen them rise, and I've seen them screw the pooch. They're looking better today and I hope it holds, but history is not on their side- yet.
 
I think people need to also see the other benefits this move could offer. It would finally bring together mobile phones / tablets and computers under a single architecture making them interchangable between OSX and iOS. You could have an iPhone running OSX as a limitied use workstation or scale it up to a laptop size unit with more power and thermal allowances. It would also allow Apple to spread out into a new market by itself and not be compared to PC's anymore hardware wise.

We also need to keep in mind these processors technically allow Apple to walk away not only from Intel but also NVIDIA and AMD as they incorporate GPU's also. The advantages are make it almost a no brainer and if anyone can do it, its Apple. Microsoft give up on shit far too quickly as soon as they hit a bump in the road.
 
aokman said:
It would finally bring together mobile phones / tablets and computers under a single architecture making them interchangable between OSX and iOS.
Click to expand...

Something to consider: we're seeing containerization and cross-platform frameworks grow to a point that the OS and hardware is becoming irrelevant. Code like Java, Go, and Python- not to mention the web languages- can be compiled and run anywhere. More important is the consideration of platform- desktop, laptop with touch, tablet, phone, watch... fridge, car, IoT?

As the frameworks evolve, expect that to be tackled too.
 
IdiotInCharge said:
That you're making poor comparisons. Especially when it comes to the outdated Intel hardware that Apple uses (for reasons inexplicable). And even if you compare Intel's current ultrabook offerings, we know they're behind on process- if you are going to bet on that situation to continue, you're betting on Intel folding. It's not a sound bet. Make honest comparisons, when I asked if you were being serious, that was honest. Comparing an Intel 28w CPU to an ARM SoC is wildly off base.
Click to expand...
Not sure why you would consider it "wildly" off base to compare the two directly, but I can agree with you on the Intel 28 watt TDP CPU being outdated tech, for sure.

This I find especially silly; you can look at 'trajectories', but you're missing that ARM (and AMD for that matter) are catching up with the low-hanging fruit; further, your denigration of x86 versus ARM is also off-base. The two are more alike than they are different, and you could strip down x86 to be power-competitive with ARM the same as you could build up ARM to be performance competitive with x86. The barriers for both are ecosystem-based. Intel x86 CPUs have shipped in phones- how many Top 500 ARM supercomputers are there? It's a poor comparison, but it illustrates the point- and I do earnestly believe that there will be ARM supercomputers in the Top 500; I just don't believe that the reason will be the instruction set.
Click to expand...
If you had just said this from the beginning, it would have made a lot more sense.
Also, most of those Top 500 supercomputers using x86-64 CPUs aren't computationally powerful because of the CPUs, it is primarily because of the GPUs being employed - not sure why this couldn't be a 'thing' with ARM64 eventually, or any ISA for that matter with enough I/O for the peripherals.

Last one: you're talking about true innovation. AMD isn't innovating except with respect to themselves; they're working on catching up to Intel. And they're still catching up to four year old Intel tech. Good for AMD! But it's not 'real innovation'. AVX512 is one example; Xeon Phi is another, if we're talking just about 'x86 innovation' and not AMD vs. Intel, because AMD seriously loses the overall competition. Their current limelight is largely made possible by an expanding market that Intel cannot fully satisfy combined with a willingness to sacrifice margins for market share- something Intel is pretty stubborn about.
Click to expand...
I agree with you on that, and that does make more sense, thanks for clarifying your point.
As for AVX512, what makes that so "innovative", though? (really want to know on this one)

Really last one: I don't give two shits about these companies' stocks. My point is that your apparent ignorance comes across as AMD worship to such a degree that it looks outright intentional. I'll give you the benefit of the doubt that that's not the case, but I'm telling you about it so you can see what you've written from that perspective.
Click to expand...
I have zero AMD worship, it's just that they are the lesser of two evils at this point and it is nice to see them make a comeback, at least in the price/performance market, after all of these years, mainly since competition is good for our wallets and innovation/competition in general (I do remember when AMD was on top circa early 2006 and offered their flagship CPU, the FX-60, for a small sum of $1200 - that quickly changed when the Intel Core 2 CPUs appeared).
By "evils" I mean x86-64, and truth be told, I absolutely cannot wait for x86-64 to officially die off - it and Intel's royalty & license stranglehold on that architecture has held the world back, computationally and technologically, for long enough; I have a sour feeling that NVIDIA is trying this same bullshit as we write this, but with no competition from any front, they have no reason to stop (might makes right, or in this case, the price).

ARM64 is the closest, competitively speaking and mainstream consumer market, ISA to make this happen.
 
Last edited:
IdiotInCharge said:
Something to consider: we're seeing containerization and cross-platform frameworks grow to a point that the OS and hardware is becoming irrelevant. Code like Java, Go, and Python- not to mention the web languages- can be compiled and run anywhere. More important is the consideration of platform- desktop, laptop with touch, tablet, phone, watch... fridge, car, IoT?

As the frameworks evolve, expect that to be tackled too.
Click to expand...
Those are really good points - why didn't you just say this from the beginning???
 
  • Like
Reactions: DocNo
like this
Red Falcon said:
Not sure why you would consider it "wildly" off base to compare the two directly, but I can agree with you on the Intel 28 watt TDP CPU being outdated tech, for sure.
Click to expand...

It's wild compared to the 15w quad-core that I'm typing this on- and Intel makes lighter-weight Core CPU than this. Nearly all are faster than shipping ARM cores.

Red Falcon said:
it is primarily because of the GPUs being employed - not sure why this couldn't be a 'thing' with ARM64 eventually, or any ISA for that matter with enough I/O for the peripherals.
Click to expand...

Agreed! The CPU in these things mostly just exists to keep the GPUs- or other computer processors- fed with data. Using a CPU for compute should only be done for very complex code.

Red Falcon said:
As for AVX512, what makes that so "innovative", though? (really want to know on this one)
Click to expand...

Considering the above, what can be done to x86? AVX512 like its SIMD forbears (SSE being the big one) extends x86 to do some of that compute without specialized processors. That's innovation.

Red Falcon said:
By "evils" I mean x86-64, and truth be told, I absolutely cannot wait for x86-64 to officially die off - it and Intel's royalty & license stranglehold on that architecture has held the world back, computationally and technologically, for long enough.
ARM64 is the closest, competitively speaking and mainstream consumer market, ISA to make this happen.
Click to expand...

It's an interesting perspective; Intel has dominated the desktop CPU market through x86, though certainly not without hard work on their part. However, it this point also brings the question: what's to stop ARM from becoming Intel, so to speak?

Obviously they produce no end products, but once they're in a commanding market position, what's to stop them from abusing it?

And I'll agree that ARM is the closest from a marketability perspective, but if the idea is to remove the proprietary stranglehold on computing, why not go actually free with something like RISC-V?
 
IdiotInCharge said:
Something to consider: we're seeing containerization and cross-platform frameworks grow to a point that the OS and hardware is becoming irrelevant. Code like Java, Go, and Python- not to mention the web languages- can be compiled and run anywhere. More important is the consideration of platform- desktop, laptop with touch, tablet, phone, watch... fridge, car, IoT?

As the frameworks evolve, expect that to be tackled too.
Click to expand...

Certainly true but to what extent do those programming languages go into local desktop applications (coding noob)?

It has been interesting to see how everyone said in the past if Apple move to ARM, software development will die and so will Apple but even with the development of the iPad Pro, we have seen amazing apps like Affinity Photo matching desktop counterparts.
 
Well, Nvidia tried a similar move where they merged the gaming/pro/mobile/phone architectures together and it didn't end well for them.

Having the same architecture across your product stack sounds nice in meeting rooms, but doesn't pan out in reality. Your product will have no competitive edge and will just be mediocre across all use cases.
 
It's so weird to knock AMD for innovation while touting x86-64 (who made the ISA) then mention the now dead and underwhelming Phi as innovation.
 
trandoanhung1991 said:
Well, Nvidia tried a similar move where they merged the gaming/pro/mobile/phone architectures together and it didn't end well for them.

Having the same architecture across your product stack sounds nice in meeting rooms, but doesn't pan out in reality. Your product will have no competitive edge and will just be mediocre across all use cases.
Click to expand...

It has been true in the past sure but do you see anything fundamentally stopping Apple from doing this and being successful? It seems like a no brainer from a performance point of view and it allows Apple to grow beyond Intel rather than being limited by PC counterparts.
 
aokman said:
Certainly true but to what extent do those programming languages go into local desktop applications (coding noob)?
Click to expand...

Note the word 'frameworks'; essentially pre-combined collections of code libraries that provide out-of-the-box functionality. One common example is Wordpress, but there are frameworks for just about anything. A game engine could be considered to be a very complicated and/or high-end framework, and note that the major engines have been developed to run on different combinations of OS and architecture. You have Unreal Engine running on ARM Android and IOS, on Windows, Linux, and Apple PC hardware, and very likely others (if you consider the consoles 'others').

aokman said:
we have seen amazing apps like Affinity Photo matching desktop counterparts.
Click to expand...

Leveraging the local compute power beyond the CPU is something Apple has really pushed for- their mobile GPUs are massive! I would however caution the use of the word 'matching'; there's the appearance of matching, and then there's empirically proven to be matching.

trandoanhung1991 said:
Well, Nvidia tried a similar move where they merged the gaming/pro/mobile/phone architectures together and it didn't end well for them.
Click to expand...

Nvidia's CPU efforts have been... poorly timed. That doesn't mean that they cannot try again, the same way Apple is- it's not like Linux lacks for software to base a system upon, and if Nvidia were to engineer in the processing power, such a system could easily rival a ultrabook.

Of course, the bigger issue- the same with Intel shipping phone SoC's and ARM server processors- is that there has to be a market for the resulting products. Apple has this built-in, whereas Nvidia and Intel really do not. So they may follow Apple, but I don't expect either of them to try to lead.
 
diehardman said:
It's so weird to knock AMD for innovation while touting x86-64 (who made the ISA) then mention the now dead and underwhelming Phi as innovation.
Click to expand...

Not really.

x86-64 wasn't technically innovative- double the length and number of registers, done. What made x86-64 innovative is that Intel refused to do it, and when AMD did it, Microsoft eventually supported them. Good gamble on AMD's part for sure, but not really technical innovation.

In contrast, Xeon Phi (and I'm not talking about the failed graphics experiment) puts dozens of out-of-order x86 CPUs on a die and stitches them in with SIMD, supporting hundreds of threads. There's an incredible flexibility there that really is innovative, but has limited applicability- like all actual innovations to x86.

IPC, clockspeed, and task energy are the real points of innovation, and while AMD is catching up, they're still behind Intel's aging architecture. The real excitement to me will be Zen 2 and Intel's post-Skylake architecture hitting in the 2019/2020 timeframe, fingers crossed for both.
 
My point is that If you list AVX512 as innovation, you can't bemoan x86-64 as.. not.

One could just as easily dismiss AVX512 as a wider AVX256 (or FMA3 really) with a couple dozen new extensions.
 
IdiotInCharge said:
post 71 - snip
Click to expand...
Those were all very good points, thank you for explaining, I will look into these further.
I appreciate your input and response a lot!

And I'll agree that ARM is the closest from a marketability perspective, but if the idea is to remove the proprietary stranglehold on computing, why not go actually free with something like RISC-V?
Click to expand...
I completely forgot about RISC-V - really need to look into how this is doing.
 
You must log in or register to reply here.
Back
Top