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

sblantipodi

2[H]4U
Joined
Aug 29, 2010
Messages
3,571
The fuck it is. Cause apple made a shitty arms laptop x86 is dying? X86 will be here for the foreseeable future.

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.
 

Keljian

[H]ard|Gawd
Joined
Nov 7, 2006
Messages
1,279
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.
You’re wrong
https://arstechnica.com/gadgets/2020/11/hands-on-with-the-apple-m1-a-seriously-fast-x86-competitor/
In specific apps in specific use cases it is faster, but not generally for compute, not yet.

People called x86 dead in the 90s, didn’t happen, people called it dead in the 2000s, then x64 came about.

It is far from dead.
 

sblantipodi

2[H]4U
Joined
Aug 29, 2010
Messages
3,571
You’re wrong
https://arstechnica.com/gadgets/2020/11/hands-on-with-the-apple-m1-a-seriously-fast-x86-competitor/
In specific apps in specific use cases it is faster, but not generally for compute, not yet.

People called x86 dead in the 90s, didn’t happen, people called it dead in the 2000s, then x64 came about.

It is far from dead.

Did you even read the article you posted?

They tested it with various scenarios and arm processors seems to be faster in every scenarios

M1 it's an octa core for an ultra portable laptop, but what can we see from a 16 cores desktop CPU?
 

daglesj

Supreme [H]ardness
Joined
May 7, 2005
Messages
5,306
I just need to cost for another 5 years or so and I can retire, dump all my IT gear and just buy a Chromebook (ARM or x86 who cares)...

Can't wait.
 
  • Like
Reactions: travm
like this

lopoetve

Fully [H]
Joined
Oct 11, 2001
Messages
30,452
The transition, if it were to happen, will take a long time. Enterprise software is written for x86 - tolerance for emulation until we have a LOT of details on how the microcode handles things / etc will take a very long time, as will rebuilding for ARM/RISC. Because it's written for x86, the development systems are written for x86. And people want to do multiple things with what they own, which is why x86 won in the first place (not many Power/SPARC/IRIX home systems, Apple aside). x86 won by being affordable, fast enough, common, and commodity. ARM has a ways to go on that outside of integrated devices (phone/tablet/laptops like Apple makes); different markets.

Not saying it CAN'T happen - there was a time that most folks thought RISC would be the winner over CISC x86 - but it's going to be a slow transition if it DOES happen.

And there's a huge market for add-on parts that would have to be developed for that ecosystem too.
 

Keljian

[H]ard|Gawd
Joined
Nov 7, 2006
Messages
1,279
Did you even read the article you posted?

They tested it with various scenarios and arm processors seems to be faster in every scenarios

M1 it's an octa core for an ultra portable laptop, but what can we see from a 16 cores desktop CPU?

I did, I also looked at the numbers...
 

zehoo

Limp Gawd
Joined
Aug 22, 2004
Messages
378
It's a great little chip, and I'll probably pick up a 2nd gen arm mac mini for music production / htpc duties, but an x86 replacement it is not.
 

Spartacus09

[H]ard|Gawd
Joined
Apr 21, 2018
Messages
1,569
Did you even read the article you posted?

They tested it with various scenarios and arm processors seems to be faster in every scenarios

M1 it's an octa core for an ultra portable laptop, but what can we see from a 16 cores desktop CPU?
Did you read the article either? It still showed x86 desktop platforms beating it handily, the Apple M1 is starting out at a 5nm design specification.
AMD are still at 7nm (soon to be 5nm in 2021) and in Intel's case 14nm, theres still room for power, performance and heat improvements before reaching the same die size.
While it is a amazing performer for the power/wattage ratio and a promising long term solution, x86 will still be relevant for years to come.
 

buttons

2[H]4U
Joined
Oct 12, 2011
Messages
2,051
I see absolutely no one switching to these chips except people that like chromebooks or running all their software through a web browser. The benchmark i saw for cinebench, showed the M1 slightly beat intel in single threaded performance and get absolutely annihilated by the Ryzen 4800u chip in multithreaded.
 

LukeTbk

Limp Gawd
Joined
Sep 10, 2020
Messages
396
I see absolutely no one switching to these chips except people that like chromebooks or running all their software through a web browser. The benchmark i saw for cinebench, showed the M1 slightly beat intel in single threaded performance and get absolutely annihilated by the Ryzen 4800u chip in multithreaded.
Cinebench R23 with native ARM support do tell a different story:
https://arstechnica.com/gadgets/2020/11/hands-on-with-the-apple-m1-a-seriously-fast-x86-competitor/

On 4 thread vs 4 thread it does not get annihilated by even a 5950x
 

Decko87

[H]ard|Gawd
Joined
Sep 23, 2007
Messages
1,831
I'm not nuts about the idea of apple having a complete monopoly on their devices. It was actually a good thing they used intel for their chips, I think this is just going to make them an even worse monolithic corporation. I don't see x86 going away due to software support. Even smaller software companies use older software which won't likely be ported to work on ARM any time soon.
 

lopoetve

Fully [H]
Joined
Oct 11, 2001
Messages
30,452
Cinebench R23 with native ARM support do tell a different story:
https://arstechnica.com/gadgets/2020/11/hands-on-with-the-apple-m1-a-seriously-fast-x86-competitor/

On 4 thread vs 4 thread it does not get annihilated by even a 5950x
But the 5950 has 4x the threads they limited it to. And was still faster even limited. So, jump that again - it's a different WORLD of performance, but also a different world of power and heat budgets.

edit: and while their 4 thread / 4 thread comparison was closer, the 5950x has 8X that number of threads available, and that's before you consider things like Threadripper. Different world still.
 

LukeTbk

Limp Gawd
Joined
Sep 10, 2020
Messages
396
But the 5950 has 4x the threads they limited it to. And was still faster even limited. So, jump that again - it's a different WORLD of performance, but also a different world of power and heat budgets.

edit: and while their 4 thread / 4 thread comparison was closer, the 5950x has 8X that number of threads available, and that's before you consider things like Threadripper. Different world still.
Not sure the but, I did explicitly say that was the case, it is still a multithread scenario and compare that work made by each of those thread.

If I render using 10 M1 in parallel (not like that would be hard to cool and power nor an hard task to parallel), is it still a different world of performance ?

To evaluate if people would want to shift one day really soon (say has soon has 2022/2023) for that type of task, looking right now the 4 thread vs 4 thread seem the way to go to have an idea, except if there is some reason that it does not scale well (hard to imagine for something like rendering).
 

DejaWiz

Fully [H]
Joined
Apr 15, 2005
Messages
20,598
If my research gleaned accurate results:
Apple's computer market share, globally, is between 7-11% in recent past quarters.

In order for "x86 to die", Apple is going to have to single-handedly:
1. ensure that compatibility with all existing MacOS software/programs is 100% perfect.
2. get all of their current users to buy ARM-based Mac computers.
3. convince the rest of the market that they need to ditch x86/x64 and go with ARM-based Mac computers, as well.

This is going to be utterly impossible, since Intel and AMD x86-based computing currently dominate the data center and corporate enterprise environments, which is the bread and butter segment of processor manufacturers.
This means Microsoft isn't going anywhere, because their server and workstation OS's aren't going anywhere.

Example: AMD's data center market share is now 10%, up from less than 1% in 2017, thanks to Ryzen, Threadripper, and Epyc.

Example 2: Microsoft, themselves, attempted ARM-based computing with the original Surface RT which ended in highly predictable fashion: it failed...miserably.

Good luck, Apple...You're entering a new chapter in the same old game, and your results will still be identical to what they always are: yawn.
 
Last edited:

LukeTbk

Limp Gawd
Joined
Sep 10, 2020
Messages
396
In order for "x86 to die", Apple is going to have to single-handedly:
1. ensure that compatibility with all existing MacOS software/programs is 100% perfect.
2. get all of their current users to buy ARM-based Mac computers.
3. convince the rest of the market that they need to ditch x86/x64 and go with ARM-based Mac computers, as well.
I am not so sure that it need that for x86 to die.

A more possible road, imo

1) Amazon cloud win that market
2) NVidia Arm and others make the next generation of console become Arm base in 2025/2026 and shift gaming to arm (if they are bound to keep an under 200watt machine, I could see arm compete at that energy enveloppe).
3) Apple Arm is a big success.
This is going to be utterly impossible, since Intel and AMD currently dominate the data center and corporate enterprise environments, which is the bread and butter segment of processor manufacturers.
But there is already crack in that domination, like Amazon data center (I think they are the biggest in that sector), isn't arm getting close above 15% of the data center market share, that could be bigger than AMD quite soon.

In networking arm is also getting really big really fast.
 

lopoetve

Fully [H]
Joined
Oct 11, 2001
Messages
30,452
Not sure the but, I did explicitly say that was the case, it is still a multithread scenario and compare that work made by each of those thread.

If I render using 10 M1 in parallel (not like that would be hard to cool and power nor an hard task to parallel), is it still a different world of performance ?

To evaluate if people would want to shift one day really soon (say has soon has 2022/2023) for that type of task, looking right now the 4 thread vs 4 thread seem the way to go to have an idea, except if there is some reason that it does not scale well (hard to imagine for something like rendering).
Then I’ll cluster 10 5950s. We’ve been doing that for a long time.
And no, it doesn’t scale well for a lot of things. Map reduce and other workloads are easy, others are hard.
Plus there’s no system with 10M1 in it anywhere.
 

lopoetve

Fully [H]
Joined
Oct 11, 2001
Messages
30,452
I am not so sure that it need that for x86 to die.

A more possible road, imo

1) Amazon cloud win that market
2) NVidia Arm and others make the next generation of console become Arm base in 2025/2026 and shift gaming to arm (if they are bound to keep an under 200watt machine, I could see arm compete at that energy enveloppe).
3) Apple Arm is a big success.

But there is already crack in that domination, like Amazon data center (I think they are the biggest in that sector), isn't arm getting close above 15% of the data center market share, that could be bigger than AMD quite soon.

In networking arm is also getting really big really fast.
Networking is low power. Most of the Amazon ARM instances are low power. Big enterprise compute is not all low power. And a lot of it is t written for ARM either.
 

DejaWiz

Fully [H]
Joined
Apr 15, 2005
Messages
20,598
I am not so sure that it need that for x86 to die.

A more possible road, imo

1) Amazon cloud win that market
2) NVidia Arm and others make the next generation of console become Arm base in 2025/2026 and shift gaming to arm (if they are bound to keep an under 200watt machine, I could see arm compete at that energy enveloppe).
3) Apple Arm is a big success.

But there is already crack in that domination, like Amazon data center (I think they are the biggest in that sector), isn't arm getting close above 15% of the data center market share, that could be bigger than AMD quite soon.

In networking arm is also getting really big really fast.

Intel's Data Center Group saw almost 36% growth in the past year, in the face of ARM-based server solutions going live (pretty much *solely* because of Amazon deploying their own Graviton processors for cloud farms).

Once Intel catches up to AMD in the node department, 10 and 7 nm Intel server parts are likely going to become another reason for more of Intel's DCG growth, because Intel has a knack for delivering some of the best performance per watt for very high density data center environments.

Let's keep in mind: While ARM is growing in the data center segment (read: Amazon), Intel and AMD aren't sitting back on their laurels...they are both aggressively innovating at a much faster pace than they were even 5 years ago, because they have those revenue and profit streams to protect.
Case in point: AMD's Zen3 architecture boasts a roughly 19% increase in IPC versus Zen2...on the same 7nm node. They already have the ball rolling to utilize 5nm.

x86 may not win Amazon back over the long run, but Amazon isn't the majority factor, what with a dozen or so other megacorporations and the thousands and thousands of smaller businesses all running x86-based data centers on the entirety of Earth.
 

GotNoRice

[H]F Junkie
Joined
Jul 11, 2001
Messages
9,836
First of all, why is this in the video card section?

2nd, this really isn't about Apple. Apple is Apple and it's users don't care about the technical details of their computers, they just want to be able to show off their latest Apple laptop or iPhone the same way others show off their jewelry. No one is thinking that Apple is going to conquer x86.

This is about ARM, and changing perceptions about ARM. If Apple can do it, then others will do it also. If ARM takes over, it will be due to chips from companies like Nvidia or Qualcomm, combined with Windows for ARM and a compatibility layer that actually works well. Or worse, we will start to see ARM-based Chrome OS desktops :sick:
 

LukeTbk

Limp Gawd
Joined
Sep 10, 2020
Messages
396
x86 may not win Amazon back over the long run, but Amazon isn't the majority factor, what with a dozen or so other megacorporations and the thousands and thousands of smaller businesses all running x86-based data centers on the entirety of Earth.
A lot of them will continue to buy cloud service instead of making their own solution, but regardless I am not sure if you disagree that it is the more likely route/scenario if Arm become the leader in most scenario has soon has say 2030 than the one you described.
 

TheSlySyl

Gawd
Joined
May 30, 2018
Messages
746
I personally want it to die. I'd like more competition than just Intel and AMD.

I also don't see it happening anytime soon. I also don't see it happening without ARM having near 100% perfect X86 emulation.

It'll eventually fade out, but I don't think it will "die" for good for at least a decade+
 

GotNoRice

[H]F Junkie
Joined
Jul 11, 2001
Messages
9,836
It's been a "dying platform" for as long as I can remember. Everything is somehow going to beat it, but never does. Scare tactics to get clicks.

If you consider mobile devices to be computers, which they are, then ARM already has far greater market share. x86 is not "dying". It will be around for a very long time, however, it has already been "beat" by ARM. For many people, give them a USB-C monitor and a bluetooth keyboard/mouse and they would be happy to use their phone as their new "desktop computer".
 

DejaWiz

Fully [H]
Joined
Apr 15, 2005
Messages
20,598
A lot of them will continue to buy cloud service instead of making their own solution, but regardless I am not sure if you disagree that it is the more likely route/scenario if Arm become the leader in most scenario has soon has say 2030 than the one you described.

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.
 

longblock454

[H]ard|Gawd
Joined
Nov 28, 2004
Messages
1,868
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.

This is an interesting point. I'll assume the M1 took significant R&D as well capitol, does it's projected market share (Mac ecosystem) support such a decision, high hopes of growth or some alternate plans we've not seen yet?
 

DejaWiz

Fully [H]
Joined
Apr 15, 2005
Messages
20,598
This is an interesting point. I'll assume the M1 took significant R&D as well capitol, does it's projected market share (Mac ecosystem) support such a decision, high hopes of growth or some alternate plans we've not seen yet?

As far as I can tell, Apple's only goal with Apple ARM is use it in everything Apple in the closed Apple ecosystem.
 

thesmokingman

Supreme [H]ardness
Joined
Nov 22, 2008
Messages
6,398
I'll assume the M1 took significant R&D as well capitol, does it's projected market share (Mac ecosystem) support such a decision, high hopes of growth or some alternate plans we've not seen yet?

I doubt that. Apple has immense experience in ARM, probably more than any other in that regard. This is a money grab and it fits like a glove with their walled garden philosophy not just from an eco system perspective but hardware. Now they maximize profits in every aspect of their product, they don't wanna be paying another 3rd party.

Also remember that Apple doesn't compete with x86 anyways so whatever the speed difference doesn't matter to the mac user anyways especially since direct comparisons are not possible.
 

longblock454

[H]ard|Gawd
Joined
Nov 28, 2004
Messages
1,868
As far as I can tell, Apple's only goal with Apple ARM is use it in everything Apple in the closed Apple ecosystem.

Perhaps Mac's ~10% market share is enough, what is that quantity? Any available rudimentary estimate as to what the savings over 3rd party chips might be?
 

DejaWiz

Fully [H]
Joined
Apr 15, 2005
Messages
20,598
Perhaps Mac's ~10% market share is enough, what is that quantity? Any available rudimentary estimate as to what the savings over 3rd party chips might be?

According to many market researchers, Apple shipped around 18 million Macs last year (and projected volume is to be around the same, this year).

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:

Apple's Own Website said:
  1. Testing conducted by Apple in October 2020 using preproduction MacBook Air systems with Apple M1 chip and 8-core GPU, as well as production 1.2GHz quad-core Intel Core i7-based MacBook Air systems, all configured with 16GB RAM and 2TB SSD. Tested with prerelease Final Cut Pro 10.5 using a 55-second clip with 4K Apple ProRes RAW media, at 4096x2160 resolution and 59.94 frames per second, transcoded to Apple ProRes 422. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Air.
  2. Testing conducted by Apple in October 2020 using preproduction Mac mini systems with Apple M1 chip, and production 3.6GHz quad-core Intel Core i3-based Mac mini systems, all configured with 16GB of RAM and 2TB SSD. Prerelease Adobe Lightroom 4.1 tested using a 28MB image. Performance tests are conducted using specific computer systems and reflect the approximate performance of Mac mini.
  3. Testing conducted by Apple in October 2020 using preproduction 13‑inch MacBook Pro systems with Apple M1 chip and 16GB of RAM measuring peak single-thread performance of workloads taken from select industry-standard benchmarks, commercial applications, and open source applications. Comparison made against the highest-performing CPUs for notebooks commercially available at the time of testing. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.
  4. Testing conducted by Apple in October 2020 using preproduction 13‑inch MacBook Pro systems with Apple M1 chip and 16GB of RAM. Multithreaded performance measured using select industry‑standard benchmarks. Comparison made against latest‑generation high‑performance notebooks commercially available at the time of testing. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.
  5. Testing conducted by Apple in October 2020 using preproduction 13‑inch MacBook Pro systems with Apple M1 chip and 16GB of RAM, as well as previous‑generation Mac notebooks. Performance measured using select industry‑standard benchmarks. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.
  6. Testing conducted by Apple in October 2020 using preproduction 13‑inch MacBook Pro systems with Apple M1 chip and 16GB of RAM using select industry-standard benchmarks. Comparison made against the highest-performing integrated GPUs for notebooks and desktops commercially available at the time of testing. Integrated GPU is defined as a GPU located on a monolithic silicon die along with a CPU and memory controller, behind a unified memory subsystem. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.
  7. Testing conducted by Apple in October 2020 using preproduction 13‑inch MacBook Pro systems with Apple M1 chip and 16GB of RAM. Performance measured using select industry‑standard benchmarks. Comparison made against latest‑generation high‑performance notebooks commercially available at the time of testing. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.
  8. Testing conducted by Apple in October 2020 using preproduction Mac mini systems with Apple M1 chip, and production 3.6GHz quad-core Intel Core i3-based Mac mini systems, all configured with 16GB of RAM and 2TB SSD. Prerelease Pixelmator Pro 2.0 Lynx tested using a 216KB image. Performance tests are conducted using specific computer systems and reflect the approximate performance of Mac mini.
  9. Testing conducted by Apple in October 2020 using preproduction 13‑inch MacBook Pro systems with Apple M1 chip, 8GB of RAM, and 512GB SSD. The wireless web test measures battery life by wirelessly browsing 25 popular websites with display brightness set to 8 clicks from bottom. The Apple TV app movie playback test measures battery life by playing back HD 1080p content with display brightness set to 8 clicks from bottom. Battery life varies by use and configuration.
  10. Testing conducted by Apple in October 2020 using preproduction MacBook Air systems with Apple M1 chip and 8-core GPU, configured with 8GB of RAM and 512GB SSD. The wireless web test measures battery life by wirelessly browsing 25 popular websites with display brightness set to 8 clicks from bottom. The Apple TV app movie playback test measures battery life by playing back HD 1080p content with display brightness set to 8 clicks from bottom. Battery life varies by use and configuration.
  11. Testing conducted by Apple in October 2020 using preproduction Mac mini systems with Apple M1 chip, and production 3.6GHz quad-core Intel Core i3-based Mac mini systems, all configured with 16GB of RAM and 2TB SSD. Tested with prerelease Logic Pro 10.6.0 with project consisting of multiple tracks, each with an Amp Designer plug-in instance applied. Individual tracks were added during playback until CPU became overloaded. Performance tests are conducted using specific computer systems and reflect the approximate performance of Mac mini.
  12. Testing conducted by Apple in October 2020 using preproduction Mac mini systems with Apple M1 chip, and production 3.6GHz quad-core Intel Core i3-based Mac mini systems with Intel Iris UHD Graphics 630, all configured with 16GB of RAM and 2TB SSD. Tested with prerelease Final Cut Pro 10.5 using a complex 2-minute project with a variety of media up to 4K resolution. Performance tests are conducted using specific computer systems and reflect the approximate performance of Mac mini.
 

Riptide_NVN

[H]ard|Gawd
Joined
Mar 1, 2005
Messages
1,871
Not saying it CAN'T happen - there was a time that most folks thought RISC would be the winner over CISC x86 - but it's going to be a slow transition if it DOES happen.
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?
 
Top