X86 is a dying plafform, will AMD switch to ARM soon?

Let's say Intel is selling their processors to Apple for an average of $150 each.

That's $2.7 billion that Apple will be keeping in-house, every year. Well, to use for paying TSMC and probably soon Samsung to make their M1 processors, anyway.

Apple systems with M1 *starting* prices:
Mac Mini (desktop, no peripherals) - $699
MacBook Air (13" screen) - $999
MacBook Pro (13"/16") - $1299/$2399


There's LOTS of fine print when it comes to Apple's claims about the M1:
https://www.anandtech.com/show/16252/mac-mini-apple-m1-tested

Actual tests, on single Core performance, it bests anything Intel can put forward.
Keep in mind though, this is a 5nm chip.

Also, I'd like real-world tests.
 
LOL!!!!!!!!!

I bet when Cyrix and Via came out someone said the same shit and yet.........here we are.
 
LOL!!!!!!!!!

I bet when Cyrix and Via came out someone said the same shit and yet.........here we are.

A big difference is that Apple won't care about datacenter performance, they will tune the chip to what they do the most, content creation, and media consumption, where it will probably excel.
 
Last edited:
It's impossible to predict what the state of the market will be like in 10+ years.

Given Intel's massive DCG growth year over year while being behind AMD's 8-ball, I'd say that the market will favor x86 for the next 3-5 years.

Apple sure as heck isn't going to see any massive gains there, unless they start selling their own ARM processors as stand-alone that can be dropped into ISA severs, without demanding (in usual Apple fashion) that they be sold as turn-key Apple systems running an Apple in-house OS.

Amazon is a bit of an exception to the rule, as they have the sheer size and revenue to take on the task of converting their own data centers utilizing their own ARM-based server processor designs. Other megacorporations don't have their own server processor subsidiary.
Amazon also had massive continual investments in x86, including custom processors from Intel. And still had. Significant portions of their infra cannot run ARM because it’s a cloud version of on premise hardware- which is x86
Ah, someone with a long memory. This. x86 has been on the way out for what, 30 years now?

I mean I'm sure it'll happen. Question is, will we be dead by then?
Yep. I’ve use SPARC, irix, power- hell, I worked on a Blue Gene/L system right when they came out.

X86 ISA is like kudzu. It just ain’t going away.
 
https://www.anandtech.com/show/16252/mac-mini-apple-m1-tested

Actual tests, on single Core performance, it bests anything Intel can put forward.
Keep in mind though, this is a 5nm chip.

Also, I'd like real-world tests.

The Tomb Raider results are impressive.

I would like to see a whole gamut of game tests at varying resolutions and settings, to find out what the ceiling is for the M1.
Problem is, there is but a fraction of PC games that are also available for Mac.

For the price of a MacBook with the full M1, I'd rather buy a laptop with a 1660Ti or better and still have access to my long-established Steam, Ubi, GOG, and Epic game libraries.
 
It's clear that X86 is a dying platform.

ARM is growing too fast for X86 to compete.
https://www.anandtech.com/show/16226/apple-silicon-m1-a14-deep-dive/4



do you think that AMD will switch to ARM soon?


Underpowered and overpriced trash is what it is. This thing is a glorified tablet just like the rest of their junk line and i say this as an IPhone owner. They can’t get close to say Razer gaming/productivity notebook performance, Apple would get slapped silly.
The most they can claim is better battery life and igpu performance vs old Intel silicon.
 
Last edited:
You mean the same way they did with PPC?

I don't think that's a fair comparison, since PPC was birthed by IBM and only survived because Apple, having agreed to throw some coin at IBM's coffers to accelerate R&D with the help of Motorola and garner exclusivity, was the sole implementer of them.

IBM retained the rights to PPC, if memory serves.
 
You're comparing apples and oranges.

x86 has a ton of big companies behind it. Intel/AMD on the hardware side, Microsoft etc on the software side.
Apple owns the whole OS and pretty much everything to make the switch.

I don't see x86 going anywhere, not even with Nvidia owning ARM.
Maybe in 10-15 years Nvidia will become like Apple, with having their own tech and OS, but I don't see that happening anytime soon or even killing off x86.
 
why would ARM die? they power more devices than x86 chips do :)

Also, IBM is still making Power chips?
Yes and no. They make power chips, which are slightly different than the PowerPC chips (similar ISA, not identical across the board). It was a joint effort until ibm couldn’t get power 6 out in time with sane power budgets.
 
Have you seen how fast ARM is growing and how fast X86 is growing?

A shitty arm laptop with 10W have the same single core performance of a ryzen 5900 and the same performance of an Intel 9700k in multithread on emulated x86 app.

Doing some maths it's not difficult to understand that arm processor will be faster then every x86 in one or two years at most.

A lot of what I have seen is that ARM is not living up to the much hyped benchmarks in real world use.

ARM is certainly growing to the point where it will be a viable alternative for some, but it is not supplanting x86 in high end desktops any time soon.

If a large enough amount of ARM systems with 16x PCIe slots become available, I'm pretty sure AMD will write ARM64 based drivers for their GPU's though.

Time will tell how that turns out.

The death of x86 has been greatly exaggerated. It will be with us for decades to come.
 
Last edited:
Trash thread, but I would love too see AMD or Intel add an fpga too there silicon. Hopefully that would allow the more costly parts of x86 to be pushed to the programmable logic leaving hopefully a nice large compute based chip as the priority for the rest of the die. Maybe even get too the point where devs can program knowing they have access to a fpga and allow optimizations as wanted.
 
Every few years someone trots out PC is dying or PC gaming is dying or x86 is dying.

Come on people.
 
Intel's process stagnation has made a lot of things look a lot better than they would otherwise. I've always been impressed with Apple's capabilities with the A series chips and how far ahead they manage to stay with iOS performance, so this is not a big surprise to me. Not sure why people are trying to take away a power / performance leap from Apple - this is very impressive by any measure. What could intel be doing if they had the process lead advantage instead of being years behind at this point is a purely theoretical question in the present.
On the other hand I think this is more of a decline in x86's importance at this point than a death knell.
 
Intel's process stagnation has made a lot of things look a lot better than they would otherwise.
On the other hand I think this is more of a decline in x86's importance at this point than a death knell.
This. Skylake came out in late 2015. Intel is still using Skylake into 2021. That's 6 years of stagnation with the process node. It is 100% a mess up at the executive leadership level with 10nm and 7nm.

A lot of these CPU designs were ready to go years ago. If it remotely followed the road map, we'd have Ice Lake equivalent in late 2017 (+10% IPC), Tiger Lake equivalent in 2018 (+10% IPC), and another tick/tock cycle (another +10% IPC) right now.
 
As was mentioned earlier, Intel pushed too hard on the process technology changing too many things at the same time. TSMC has show a more conservative approach of enhancement and gradual node steps work best instead of trying to "hit it out of the park" . Intel's problem even after the 10nm miss is they had no backup plan, either.
 
I thought that ARM was going to take over x86 as well, but I don't think it will. I think that ARM and x86 will continue to be segmented.

ARM is very good in lower power scenarios, but it is only able to accomplish this because everything is tightly integrated. In the Apple M1, even the RAM is part of the main chip. Therefore, nothing is interchangeable or upgradeable.
In x86, all the parts are separate and talk to each other via interfaces, which cost power and efficiency. However, x86 is a much more scalable platform.

Take the good and bad of each platform... you will never have both unless you give up something.
 
As was mentioned earlier, Intel pushed too hard on the process technology changing too many things at the same time. TSMC has show a more conservative approach of enhancement and gradual node steps work best instead of trying to "hit it out of the park" . Intel's problem even after the 10nm miss is they had no backup plan, either.
That was definitely not the problem with Intel 10nm. It was executive leadership plain and simple.
 

X86 is a dying platform​

giphy-downsized-large.gif
 
That was definitely not the problem with Intel 10nm. It was executive leadership plain and simple.
Your response vastly oversimplifies the issues and Ignorant at best.
 
Last edited:
Your response vastly oversimplifies the issues and Ignorant at best.
Speaking with friends who worked as process engineers in TMG, it is not an oversimplification.

Executive leadership within TMG was fostering an incredibly reckless culture not dissimilar to Wall Street during the financial boom. Managers/Senior Engineers were being promoted based on "achieving" unrealistic short term goals by delivering misrepresented (or sometimes even fabricated) results. At the same time, their peers who were trying to speak up about potential issues were being retaliated against. Of course when the SHTF (or right before), those same bad managers/engineers either ducked out "seeking other opportunities" or disappeared.

And coincidentally, that's why Intel corporate training in the 2018/2019 time period conveniently started focusing on assuring workers that it is not okay to be retaliated against for bringing up "potential issues". That it should be encouraged to inform your management chain or even HR about "potential issues" if you have the data to back up said claims.
 
Speaking with friends who worked as process engineers in TMG, it is not an oversimplification.

Executive leadership within TMG was fostering an incredibly reckless culture not dissimilar to Wall Street during the financial boom. Managers/Senior Engineers were being promoted based on "achieving" unrealistic short term goals by delivering misrepresented (or sometimes even fabricated) results. At the same time, their peers who were trying to speak up about potential issues were being retaliated against. Of course when the SHTF (or right before), those same bad managers/engineers either ducked out "seeking other opportunities" or disappeared.

And coincidentally, that's why Intel corporate training in the 2018/2019 time period conveniently started focusing on assuring workers that it is not okay to be retaliated against for bringing up "potential issues". That it should be encouraged to inform your management chain or even HR about "potential issues" if you have the data to back up said claims.

Cool Story Bro. Sounds like every job I had in Corporate America for the past several decades. Thank god its over with. Yeah it's the same over at Boeing too.

Intel's lofty over reaching goals for the 10nm process node and it's failures are well documented. It's not JUST management, those issues exist EVERYWHERE.
 
Cool Story Bro. Sounds like every job I had in Corporate America for the past several decades. Thank god its over with. Yeah it's the same over at Boeing too.

Intel's lofty over reaching goals for the 10nm process node and it's failures are well documented. It's not JUST management, those issues exist EVERYWHERE.

Strange, somehow AMD appears to be executing well.

If you're talking lofty goals wise, 22nm should have been far more difficult, given the current technology at the time.
 
Strange, somehow AMD appears to be executing well.
Imagine that, have TSMC make your chips and you'll immediately get 7nm capability!

All jokes aside I give props to Intel for sticking it out this long manufacturing their own chips.
They are definitely lagging behind though, iirc they are having to consider abandoning manufacturing their own (though as a contingency plan).

Intel is still designing good CPUs. It's just to bad they are having to scale them up from 10 to 14+++++++.
Thats the name of the game nowadays, you design your own stuff still but have TSMC fabricate it ~ amd, intel, apple, qualcom, mediatek, broadcom
 
Intel is still designing good CPUs. It's just to bad they are having to scale them up from 10 to 14+++++++.
Ok thanks for agreeing with me.

It's 100% the problem of the executive leadership of TMG.
 
Theres far too much software on x86/64 for it to just be forgotten.
Whatever takes its place in PCs will need to emulate x86/64 with great performance, the only way I reckon.
I dont think its going anywhere for quite some time yet.

How AMD and Intel will both be able to use the same instruction set wont be a simple thing to resolve.
 
Last edited:
Man, if the exceptionally crappy x86 ISA was going to be defeated, it would have been already by Motorola's 68000-series CPUs back in the 1980s. No one stuck with x86 all those years ago because of it's "elegant architecture" - because x86 is a bastard no matter how you look at it. No, x86 persists for the same reasons it has always persisted... the long tail of software written for it. Once IBM chose it for their brand new PC platform, forced Intel to license it to other manufacturers to insure multi-pronged supply, and successfully made the PC platform the dominant computing platform (in conjunction with the rise of PC clones courtesy of MS-DOS and Compaq's reverse engineering of the IBM BIOS) x86 was always going to be with us in one form or another for a very long time.

Just about every architecture x86 dominance has prematurely knifed in the back has been better than x86 in just about every conceivable metric other than that same long tail of software compatibility.
 
Back
Top