Future of x86

[Pc6777]

3 hours ago, whm1974 said:

I really think that it is Marketing Nonsense by Marketeers. The Ideal that RISC ISAs will utterly replace x86 ISA based CPUs, well the that boat has already Sailed and Sunk.

 

Anyone remember the Advance RISC Computer? This was supposed to be a Common Platform for MIPS ISA based computers to replace the IA32 platform. It was to do the same with the infamous x86 BIOS. Microsoft was one of the founding members.

 

One of reasons it sunk is due to Intel always improving both x86 and their FAB Nodes. And of course both AMD and Cycrix provided the Fire under Intel as well.

Im a big fan of backwards compatibility, if they made a backwards compatible cpu that's not x86, but could run x86 operating systems and programs I would be cool with that, but if arm comes in trying to kill x86 and my x86 programs then I would really dislike the change. Im all about the future and innovation is good, but not when you destroy the great relics of the past. 

[Fnige]

I wonder if something thats compatible with both well would actually win for the future of CPUs

[Pc6777]

38 minutes ago, SlimyPython said:

I wonder if something thats compatible with both well would actually win for the future of CPUs

probably, isnt that what happened in the early 2000s? the design that was compatible with the new and the old won. if there's full support for x86 in bios and everything im all in, if not I will never switch, 

[papajo]

3 hours ago, Pc6777 said:

probably, isnt that what happened in the early 2000s? the design that was compatible with the new and the old won. if there's full support for x86 in bios and everything im all in, if not I will never switch, 

Since the 80s computers use x86, bios has nothing to do with compatibility regarding the processor's architecture. It is a matter of how the processor works internally (how it understands and executes code produced by a program to put it simply) and since the mid 70s CISC CPUs were the norm (mainly because of the  intel 8080 chip which was the father of the first x86 CPU intel 8086) 

[Pc6777]

2 hours ago, papajo said:

Since the 80s computers use x86, bios has nothing to do with compatibility regarding the processor's architecture. It is a matter of how the processor works internally (how it understands and executes code produced by a program to put it simply) and since the mid 70s CISC CPUs were the norm (mainly because of the  intel 8080 chip which was the father of the first x86 CPU intel 8086) 

I remember, I was thinking of the transition from 32 to 64 bit, and I thought intel dropped support for windows 7 on there new chipsests I read that somewhere, could be false. That's why I assumed bios had something to do with compatibility.

[Pc6777]

2 hours ago, papajo said:

Since the 80s computers use x86, bios has nothing to do with compatibility regarding the processor's architecture. It is a matter of how the processor works internally (how it understands and executes code produced by a program to put it simply) and since the mid 70s CISC CPUs were the norm (mainly because of the  intel 8080 chip which was the father of the first x86 CPU intel 8086) 

I remember, I was thinking of the transition from 32 to 64 bit, and I thought intel dropped support for windows 7 on there new chipsests I read that somewhere, could be false. That's why I assumed bios had something to do with compatibility. This might be slightly off topic but I don't want to start a new thread for this because it does have to do with this topic, how do old operating systems work on new hardware. Like does xp still work on some modern stuff. How good is windows 7 compatibility with modern hardware? Will the current build of windows 10 work with new hardware in 10-15 years assuming it's still x86? I have lots of oldish and new isos that I save in case there's compatibility problems or I need to do stuff offline

[Murasaki]

just like HDDs - not soon enough to worry about it 

[whm1974]

6 hours ago, Pc6777 said:

I remember, I was thinking of the transition from 32 to 64 bit, and I thought intel dropped support for windows 7 on there new chipsests I read that somewhere, could be false. That's why I assumed bios had something to do with compatibility. This might be slightly off topic but I don't want to start a new thread for this because it does have to do with this topic, how do old operating systems work on new hardware. Like does xp still work on some modern stuff. How good is windows 7 compatibility with modern hardware? Will the current build of windows 10 work with new hardware in 10-15 years assuming it's still x86? I have lots of oldish and new isos that I save in case there's compatibility problems or I need to do stuff offline

Consider that most business skip Windows 8 and people like me used a Third party Start Menu replacement if they did install the OS. Intel had to provide drivers for new hardware to support Win7 until fairly recently.

 

 

[Commodus]

To answer the original question:

 

Odds are x86 still has a long while to go. However, we're also on the cusp of a transformation that may not be particularly kind to x86 and could relegate it to a has-been if its chief backers (Intel, AMD and Microsoft) aren't careful.

 

Apple's M1 isn't universally faster than equivalent x86 chips, but it shows what's possible and where the industry is going: fast, highly responsive ARM chips that last seemingly forever on battery. While x86 has gotten better, it's struggling on certain fronts. And remember, this is the first generation with Apple's lowest-end chip — there's a real chance future chips will take an outright lead.

 

Windows on ARM currently sucks, but that's more a problem for Microsoft than ARM.

 

I'm not so naive as to think ARM and Apple will dominate the market in the near future, or that Windows inertia won't keep people tied to x86 for a while. But you can see an upset coming, and the Intel/AMD/Microsoft nightmare is a scenario where Apple and maybe Google's Chrome OS partners are offering computers that completely run rings around x86 PCs.

[Pc6777]

1 hour ago, Commodus said:

To answer the original question:

 

Odds are x86 still has a long while to go. However, we're also on the cusp of a transformation that may not be particularly kind to x86 and could relegate it to a has-been if its chief backers (Intel, AMD and Microsoft) aren't careful.

 

Apple's M1 isn't universally faster than equivalent x86 chips, but it shows what's possible and where the industry is going: fast, highly responsive ARM chips that last seemingly forever on battery. While x86 has gotten better, it's struggling on certain fronts. And remember, this is the first generation with Apple's lowest-end chip — there's a real chance future chips will take an outright lead.

 

Windows on ARM currently sucks, but that's more a problem for Microsoft than ARM.

 

I'm not so naive as to think ARM and Apple will dominate the market in the near future, or that Windows inertia won't keep people tied to x86 for a while. But you can see an upset coming, and the Intel/AMD/Microsoft nightmare is a scenario where Apple and maybe Google's Chrome OS partners are offering computers that completely run rings around x86 PCs.

maybe they will still make them in the future even if they are slower, for legacy stuff, for a while at least. 

[Pc6777]

3 hours ago, whm1974 said:

Consider that most business skip Windows 8 and people like me used a Third party Start Menu replacement if they did install the OS. Intel had to provide drivers for new hardware to support Win7 until fairly recently.

 

 

is it still possible top put windows 7 on new hardware with some tweaks? and how did they drop support, and why?

[whm1974]

2 minutes ago, Pc6777 said:

is it still possible top put windows 7 on new hardware with some tweaks? and how did they drop support, and why?

Microsoft recently dropped support in in order to support Windows 10 better. Plus Win7 came out in 2009.

[Pc6777]

28 minutes ago, whm1974 said:

Microsoft recently dropped support in in order to support Windows 10 better. Plus Win7 came out in 2009.

you mean security updates? you could put windows 7 on new hardware tho, just with no updates?

[papajo]

On 2/8/2021 at 11:59 AM, Murasaki said:

just like HDDs - not soon enough to worry about it 

but unlike HDDs CISC x86 has no fear of being obsolete by something else ... this entire topic is based on memes and marketing (probably from Apple M1 ) 

 

They are not as fast as x86 as solid and redundant as safe  and never will be. 

 

On 2/8/2021 at 9:49 PM, Pc6777 said:

is it still possible top put windows 7 on new hardware with some tweaks? and how did they drop support, and why?

they dropped support in the sense that they wont help you with any issue, they wont release any new updates for safety or compatibility or whatever.

 

The OS it self can be installed to your system and run 

 

On 2/8/2021 at 8:28 PM, Commodus said:

Apple's M1 isn't universally faster than equivalent x86 chips, but it shows what's possible and where the industry is going: fast, highly responsive ARM chips that last seemingly forever on battery. While x86 has gotten better, it's struggling on certain fronts. And remember, this is the first generation with Apple's lowest-end chip — there's a real chance future chips will take an outright lead.

No it shows how fast can things go if you are a closed software ecosystem and work on your own hardware 

 

They are faster (compared to older i5 found in older MacBooks ) because a) code efficiency  b) bad thermals in said macbooks c) compared to old stuff.

 

 

If you compare it to a  decent  portable x86 cpu implementation it will either outright lose 

 

Or in instances it will win it will be because the apple OS and code being better optimized  than whatever substitute you run to test the x86CPU 

[Commodus]

1 hour ago, papajo said:

No it shows how fast can things go if you are a closed software ecosystem and work on your own hardware 

 

They are faster (compared to older i5 found in older MacBooks ) because a) code efficiency  b) bad thermals in said macbooks c) compared to old stuff.

 

 

If you compare it to a  decent  portable x86 cpu implementation it will either outright lose 

 

Or in instances it will win it will be because the apple OS and code being better optimized  than whatever substitute you run to test the x86CPU 

Never mind that the claims of outright losses or software optimizations aren't entirely true... your argument has that "but apart from that, Mrs. Lincoln, how was the play?" whiff to it.

 

If the Mac is faster, however it accomplishes that goal, it's faster. People buy computers to get things done, not to obsess over how a chip would perform in an absolutely neutral environment. It's good to have a variety of tests that expose possible weaknesses, but if tight integration between the OS and hardware leads to an advantage without any kind of cheating, it's worth noting.

[papajo]

11 minutes ago, Commodus said:

If the Mac is faster, however it accomplishes that goal, it's faster. People buy computers to get things done, not to obsess over how a chip would perform in an absolutely neutral environment

You realize that you participating in a topic with title "What is the future of x86" right?  not a "mac vs pc" kind of topic.

 

Also I didnt say that the mac is faster I said that in the few occasions a x86 laptop cpu doesnt squash to death  an m1 macbook is because of apples ecosystem which is optimized directly for their hardware (as where e.g 3rd party software outside apples ecosystem has to maintain compatibility with a lots of different hardware with everything that means in terms of budget limitations and software efficiency) 

 

 

[Guest]

X86 Gone is 15 years? Super very unlikely. It'll be just taking off from 2D to 3D by then, possibly mainstream. 

https://www.tsmc.com/english/news-events/blog-article-20200803

ARM can scale up, maybe we'll see it in the next XBOX? 

[Pc6777]

6 hours ago, papajo said:

You realize that you participating in a topic with title "What is the future of x86" right?  not a "mac vs pc" kind of topic.

 

Also I didnt say that the mac is faster I said that in the few occasions a x86 laptop cpu doesnt squash to death  an m1 macbook is because of apples ecosystem which is optimized directly for their hardware (as where e.g 3rd party software outside apples ecosystem has to maintain compatibility with a lots of different hardware with everything that means in terms of budget limitations and software efficiency) 

 

 

Well having a closed ecosystem directly optimized for hardware does help a lot, how do you think first party switch games look suprisgly decent while running a PC with specs equivalent to a 2007 high end rig. Still, I would rather have worse efficiency and just brute force it with better parts than be stuck in a closed ecosystem with tim cook as my overlord over all the stuff I "own". It's like comparing the speed of a train, which has a closed linear path, to the speed of a car which can go anywhere, sure the train is faster and more efficient but you can't go anywhere you want and can only stop at certain places plus the random searches and cameras (drm). Anyway, I doubt third party stuff runs as well on arm on those apple cpus, well because emmulation will always be lossy to some degree and will never run as well unless it's fully ported or made to run on arm. For once I hope intel succeedes and wins (as well as amd) to keep x86 as relevant as long as possible. The real solution is a versatile hybrid that's optimized to run arm and x86 so it's the past and the future in one chip, but I don't know anything about cpu architecture so it's probably easier said than done. I bet Microsoft is already thinking of making a separate windows 10 for arm or making windows 10 fully cross compatible and optimized for x86 and arm.

[whm1974]

5 hours ago, ShrimpBrime said:

X86 Gone is 15 years? Super very unlikely. It'll be just taking off from 2D to 3D by then, possibly mainstream. 

https://www.tsmc.com/english/news-events/blog-article-20200803

ARM can scale up, maybe we'll see it in the next XBOX? 

So can x86-64 BTW.

[LAwLz]

On 2/4/2021 at 4:24 AM, RONOTHAN## said:

By my guess, it'll be something around 10-15 years before ARM makes a stay in the higher end desktop market. Part of the reason is that x86 is an absolute work horse of an instruction set. I forget who said it but the best analogy I've heard is this: "ARM is the friend who comes over to help you move and you tell him 'move this here' then wait till he's done and then tell him 'move that there' and so one. x86 is the friend who you can tell 'move this here, move that there, move that thing over there, etc.' and he'll do it without any issue." x86 might not be as efficient, or as performant as ARM, but it makes up for it in sure brute force of running code, as ARM needs code optimized for it to run, or at least special hardware designed to accelerate it to make it usable, x86 can just let it run. In addition, with the way AMD and Intel have been leapfrogging one another in performance, the IPC improvements are starting to show that x86 might outpace the improvements of ARM. While ARM may fully take over laptops, the odds of it taking over desktop any time soon are relatively low.

This is incredibly wrong.

 

ARM does not need "code optimized for it to run" any more or less than x86 needs "code optimized for it to run".

ARM does not need special hardware designed to accelerate things any more or less than x86 needs it.

The improvements from AMD and Intel do not indicate that it might outpace improvements from ARM either. Both ARM processors (at least from Apple) and x86 are improving quite quickly right now but I certainly wouldn't say one is improving faster than the other. 

 

 

I do agree that it probably won't take over the desktop anytime soon. At least not for Windows desktops. That's not the fault of the ARM ISA however. It is because Microsoft and Qualcomm both shat their pants when it came to making compelling ARM based products.

If Microsoft and Qualcomm had released something as good as the M1 based Macs, then I think ARM on laptops and on desktops would have had a pretty decent chance. As it stands right now, Windows itself might die out before x86 loses to ARM on PCs.

[LAwLz]

15 hours ago, papajo said:

Also I didnt say that the mac is faster I said that in the few occasions a x86 laptop cpu doesnt squash to death  an m1 macbook is because of apples ecosystem which is optimized directly for their hardware (as where e.g 3rd party software outside apples ecosystem has to maintain compatibility with a lots of different hardware with everything that means in terms of budget limitations and software efficiency) 

[Citation Needed]

 

People have been saying this for ages now and it has not even once been proven to be true.

Even when people were comparing iPhones with an extremely powerful CPU and GPU to low end Android phones using terrible cores and tiny GPUs people still went "the only reason why the iPhone is faster is because of software optimization". Like no you idiot, it is faster because its SoC has literally 4 times as many transistors in it.

 

People don't write iOS (or even most MacOS) software in assembly. The performance that come out of Apple computers aren't software magic. It's backed by really good hardware.

You can see this when you compare assembly code running on both systems, for example Dav1d:

 

Also, please note that the single threaded performance run on the M1 is incorrect. It should read 91 FPS as noted by Janne:

[FakeFBI]

On 2/3/2021 at 9:24 PM, RONOTHAN## said:

By my guess, it'll be something around 10-15 years before ARM makes a stay in the higher end desktop market. Part of the reason is that x86 is an absolute work horse of an instruction set. I forget who said it but the best analogy I've heard is this: "ARM is the friend who comes over to help you move and you tell him 'move this here' then wait till he's done and then tell him 'move that there' and so one. x86 is the friend who you can tell 'move this here, move that there, move that thing over there, etc.' and he'll do it without any issue." x86 might not be as efficient, or as performant as ARM, but it makes up for it in sure brute force of running code, as ARM needs code optimized for it to run, or at least special hardware designed to accelerate it to make it usable, x86 can just let it run. In addition, with the way AMD and Intel have been leapfrogging one another in performance, the IPC improvements are starting to show that x86 might outpace the improvements of ARM. While ARM may fully take over laptops, the odds of it taking over desktop any time soon are relatively low.

arm is a low power powerful guy, but withshit tools because manny programs cant run, the x86 has been mastered and is basiclly a black belt, only getting beat a few times by the ARM

[Commodus]

5 minutes ago, ComputerBuilder said:

arm is a low power powerful guy, but withshit tools because manny programs cant run, the x86 has been mastered and is basiclly a black belt, only getting beat a few times by the ARM

That's not actually true. We've seen M1-based Macs trounce comparable x86 PCs, and even some ostensibly more powerful ones, in a variety of real-world tests. And those apps that don't run aren't limited by ARM, they're typically limited by having to run highly x86-specific code (like virtual machines).

 

Maybe you're just used to the shoddy state of Windows on ARM, where few things run natively and emulation is poor... but that doesn't represent ARM as a whole.

[Vishera]

It all depends on performance and software support,

It's possible that X86 has more than 50 years left,and it's possible that it only has 10 years left.

[Radium_Angel]

Our existing multi-core, multi-Ghz CPUs have an incredible amount of power at our fingertips. More than we could possibly know what to do with.

The trouble is not the hardware, it's how it is utilized by the OS.

 

Hell, BeOS, in the 90s, showed what was possible with a clean sheet design on single core Mhz systems that was simply mind blowing.

Windows (and the ever-willing-to-copycat Linux) is a bloated morass of legacy code, telemetry, pointless visuals and other who-knows-what-crap-lurks-beneath-the-surface, all which serve to destroy any silicon advances made by x86.