Click Here
home features news forums classifieds faqs links search
6071 members 
Amiga Q&A /  Free for All /  Emulation /  Gaming / (Latest Posts)
Login

Nickname

Password

Lost Password?

Don't have an account yet?
Register now!

Support Amigaworld.net
Your support is needed and is appreciated as Amigaworld.net is primarily dependent upon the support of its users.
Donate

Menu
Main sections
Home
Features
News
Forums
Classifieds
Links
Downloads
Extras
OS4 Zone
IRC Network
AmigaWorld Radio
Newsfeed
Top Members
Amiga Dealers
Information
About Us
FAQs
Advertise
Polls
Terms of Service
Search

IRC Channel
Server: irc.amigaworld.net
Ports: 1024,5555, 6665-6669
SSL port: 6697
Channel: #Amigaworld
Channel Policy and Guidelines

Who's Online
41 crawler(s) on-line.
 30 guest(s) on-line.
 1 member(s) on-line.


 Karlos

You are an anonymous user.
Register Now!
 Karlos:  1 min ago
 pixie:  9 mins ago
 Blade2019NL:  9 mins ago
 amigakit:  50 mins ago
 clint:  1 hr 1 min ago
 zipper:  1 hr 21 mins ago
 A1200:  2 hrs 16 mins ago
 Templario:  2 hrs 22 mins ago
 Frank:  2 hrs 44 mins ago
 MagicSN:  2 hrs 46 mins ago

/  Forum Index
   /  General Technology (No Console Threads)
      /  Why x86 doesn't need to die
Register To Post

Goto page ( 1 | 2 Next Page )
PosterThread
AmigaNoob 
Why x86 doesn't need to die
Posted on 1-Apr-2024 9:08:25
#1 ]
Member
Joined: 14-Oct-2021
Posts: 15
From: Unknown

Interesting read on RISC vs. CISC
https://chipsandcheese.com/2024/03/27/why-x86-doesnt-need-to-die/

 Status: Offline
Profile     Report this post  
hardwaretech 
Re: Why x86 doesn't need to die
Posted on 4-Apr-2024 21:19:35
#2 ]
Member
Joined: 5-May-2010
Posts: 63
From: blaine minnesota usa

Windows is already on Risc see

https://www.youtube.com/watch?v=iryLKvhsgrU
The chip gives the M3 a run for the money

 Status: Offline
Profile     Report this post  
DiscreetFX 
Re: Why x86 doesn't need to die
Posted on 5-Apr-2024 0:24:52
#3 ]
Elite Member
Joined: 12-Feb-2003
Posts: 2518
From: Chicago, IL

@hardwaretech

Are you trying to bore us to death?

_________________
Sent from my Quantum Computer.

 Status: Offline
Profile     Report this post  
agami 
Re: Why x86 doesn't need to die
Posted on 5-Apr-2024 8:50:31
#4 ]
Super Member
Joined: 30-Jun-2008
Posts: 1716
From: Melbourne, Australia

@thread

For my personal needs, I've stopped caring about the silicon a long while back.
I generally ask these questions:
1. Do the apps I like using run on it, via my choice of OS for low friction workflows?
2. Is it good value for money?
3. Can I build my own personal computer, and/or upgrade it with different components?

I needed to let go of that 3rd one when I got an M2 Mac mini last year.

I'd have no issues one day building myself a personal computer based on a socketed ARM64 CPU. Until then, x86-64 still covers questions 2 and 3, even if Windows 11 and Linux/Gnome don't quite cover question 1.


_________________
All the way, with 68k

 Status: Offline
Profile     Report this post  
fishy_fis 
Re: Why x86 doesn't need to die
Posted on 5-Apr-2024 17:09:22
#5 ]
Elite Member
Joined: 29-Mar-2004
Posts: 2163
From: Australia

Interesting times from a hardware perspective at the moment.
Both ARM and x86-64 are making very rapid improvements.
ARM is catching up at a considerable clip and x86 has seen the biggest, most rapid advancements it has in well over a decade.

Much larger gap between x86/ARM and ppc than there was between 68k and PPC when the A1's 1st came out.

I enjoy retro as much as I do modern hardware, so I really dont care that it's the case, but everything Amiga has fallen deep into the realm of retro these days. The gap between the absolute most powerful ppc hardware available to amiga/amiga-noids is astronomical.
We're back to a scenario where high end hardware can emulate ppc faster than anything native, and thats with there thus far being nothing even as close to as optimized as 68k emulation.

Feels like there's a bit of a rush to reach the "finish line" currently (with "finishing line" being having enough raw grunt as to exceed anything a person may want to do in a personal computer).

Interesting times.
Zen5 rumors are suggesting it will be the biggest jump gen on gen in x86-64 history as well.

 Status: Offline
Profile     Report this post  
ppcamiga1 
Re: Why x86 doesn't need to die
Posted on 12-Apr-2024 8:53:17
#6 ]
Cult Member
Joined: 23-Aug-2015
Posts: 829
From: Unknown

@fishy_fis

what you wrot is usuall anti powerpc propaganda bs

in reality

after 20 years first amiga one (600 MHz G3)

was and still is at least 10 times faster than any real 68k

my i9 is maybee 3 or 3 times faster than my g4 in normal mode

before fans start and my laptop try to took off

affordable arm computers (rpi) are stil slower than G4

ppc emulation on my i9 in normal mode is still 5 times slower than my g4

68k emulation is as fast as it was 20 years ago

and as it was 20 years ago still nothing "amiga" worth use on x86/arm

still nothing as good as windows

so no reasons to switch from ppc


 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Why x86 doesn't need to die
Posted on 12-Apr-2024 17:48:06
#7 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12856
From: Norway

@ppcamiga1

Its clear that where LE CPU x86 suck right now is on the FPU emulation, that’s one thing that remains for sure.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Why x86 doesn't need to die
Posted on 12-Apr-2024 18:28:14
#8 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12856
From: Norway

@DiscreetFX

Who cares if replaced or not.

There is no AmigaOS software using x86 cpu’s.
Microsoft and Apple have money and resources to switch to any CPU they like, at the same time more and more of Microsoft product is cloud bases, and web technology. So from business application point of view it does not matter what OS or CPU you have, Liniux, MacOS, Windows.

We can use AmigaOS even if was not for lack of security built into it. well as long as no one uses it for anything important, its as safe as the software you trust, and that case with any OS I suppose.

The transition to web technical good for AmigaOS users, but it also has its problems, regarding out lack of JS JIT compilers, and JS Machine Code, interpreter. Without that it always run sh*t slow, but also web browser on Windows and Linux used 2D/3D acceleration for rendering pages, the biggest issue is amount memory need to web for browsing.

Last edited by NutsAboutAmiga on 12-Apr-2024 at 06:33 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
Karlos 
Re: Why x86 doesn't need to die
Posted on 12-Apr-2024 19:30:49
#9 ]
Elite Member
Joined: 24-Aug-2003
Posts: 4491
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

x86 will never die. There are countless embedded systems running it. It's had hundreds of millions of users worldwide and even if only a diminishingly small fraction of those are passionate about it, it will have more than enough users to keep it alive in virtuo, no matter what the hardware changes to.

It was never my cup of tea as an architecture, but rightly or wrongly it changed the world.

_________________
Doing stupid things for fun...

 Status: Online!
Profile     Report this post  
NutsAboutAmiga 
Re: Why x86 doesn't need to die
Posted on 12-Apr-2024 20:18:14
#10 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12856
From: Norway

@Karlos

Just write a JIT compiler and all software will run, in emulation box, vm or whatever.
the instructionset can die, and no one will notice.

And there is no reason why they can't clean up the instruction set, and remove all the old crap from it that no one uses anymore. I image if something written for 8088, 286, or 386, its not going be software with high hardware demands.

Last edited by NutsAboutAmiga on 12-Apr-2024 at 08:22 PM.
Last edited by NutsAboutAmiga on 12-Apr-2024 at 08:21 PM.
Last edited by NutsAboutAmiga on 12-Apr-2024 at 08:19 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
BigD 
Re: Why x86 doesn't need to die
Posted on 12-Apr-2024 21:10:30
#11 ]
Elite Member
Joined: 11-Aug-2005
Posts: 7349
From: UK

@Karlos

Well it made my Mac a hell of a lot more fun and usable! Bootcamp allowed gaming on Macs to continue for a few years when Apple gave up supporting OpenGL! Elite Dangerous, Sonic Mania, Streets of Rage 4...I owe it all to x86-64 and being able to dual boot into Windows!

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios

 Status: Offline
Profile     Report this post  
Karlos 
Re: Why x86 doesn't need to die
Posted on 13-Apr-2024 0:38:56
#12 ]
Elite Member
Joined: 24-Aug-2003
Posts: 4491
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@NutsAboutAmiga

That's what I'm saying - "in virtuo", meaning that like 68K, x86 (as in 32-bit) can be emulated in software.

To be honest, though, given how x86/64 ISA is actually executed in real silicon, there probably is no reason at all to get rid of it anyway. It's all translated down into micro-architectural instructions. I don't think there are any implementations in 20+ years that directly execute the ISA.

_________________
Doing stupid things for fun...

 Status: Online!
Profile     Report this post  
agami 
Re: Why x86 doesn't need to die
Posted on 13-Apr-2024 3:11:58
#13 ]
Super Member
Joined: 30-Jun-2008
Posts: 1716
From: Melbourne, Australia

@

Quote:
i9 is maybee 3 or 3 times faster than my g4

Or maybe it's a number between 3 and 3

You don't actually have an i9. That's just arbitrary numbering for those obsessed with measuring their small male parts.

Both Raspberry Pi 4 and 5 are faster than any system based on Motorola's/Freescale's MPC7400 series CPU, not to mention more affordable.

You know what feeling I get when I boot up Mac OS X 10.5 (Leopard) on my Power Mac G5 Quad 2.5GHz, or MorphOS 3.18 on my Mac mini G4 1.5GHz?

Nothing.

No special warm and fuzzy feelings. There's no stirring in my loins. I don't see big endian byte ordering when I close my eyes. The PowerPC geek amplification field has no affect on me. Not since 2005.
I guess one could say that more pragmatic things occupy my geeky interests.

Also, unless one is a masochist, there's every reason to switch from PowerPC computing to pretty much anything else.

_________________
All the way, with 68k

 Status: Offline
Profile     Report this post  
matthey 
Re: Why x86 doesn't need to die
Posted on 13-Apr-2024 3:56:33
#14 ]
Elite Member
Joined: 14-Mar-2007
Posts: 2146
From: Kansas

Karlos Quote:

x86 will never die. There are countless embedded systems running it. It's had hundreds of millions of users worldwide and even if only a diminishingly small fraction of those are passionate about it, it will have more than enough users to keep it alive in virtuo, no matter what the hardware changes to.


There is no doubt that x86 will be emulated for a long time after it is dead for desktop and game applications. However, I don't think this is true for embedded use which will be ported to other architectures despite any difficulty. The 68k, ColdFire and SuperH were more popular than x86 and they are being replaced by ARM despite the difficulty, including a transition to little endian. Developers may wish there was big endian 68k like hardware available but despite the ease of using an emulator, hardware costs more, uses more power, is lower performance, has increased latency (interrupts are slower), etc. Emulation is the sign of a dead architecture. It may be acceptable for retro compatibility but it is not competitive for most computer uses.

Karlos Quote:

That's what I'm saying - "in virtuo", meaning that like 68K, x86 (as in 32-bit) can be emulated in software.

To be honest, though, given how x86/64 ISA is actually executed in real silicon, there probably is no reason at all to get rid of it anyway. It's all translated down into micro-architectural instructions. I don't think there are any implementations in 20+ years that directly execute the ISA.


The x86-64 ISA targets high performance but is too fat to be efficient when scaled down even to medium performance markets to compete with ARM64. ARM64 is also too fat to scale down far enough to cover the low end embedded market.

ARM
low: Thumb
mid: AArch64
high: AArch64 uoped

x86-64
low: 808x (defunct)
mid: x86 (defunct, failed attempts to bring back and use non uoped x86 and x86-64 with Atom)
high: x86-64 uoped

At one time, the 68k scaled from low to high with one architecture.

68k
low: 68k (68k created the 32 bit embedded category and led it into at least late 1990s)
med: 68k (68k had success in the desktop market but failed to dominate)
high: 68k (68k created the microprocessor workstation market)

The semi-modern in-order 68060 was more efficient than the in-order P5 Pentium which Intel used in attempts to scale down with the in-order Atom. The 68k has better code density than x86, x86-64 and AArch64. It is even comparable to Thumb which scales very low for embedded use. There are no OoO uoped 68k CPU cores so maybe the 68k wouldn't scale as high as x86-64 but the 68k has many of the CISC performance traits of x86-64 which give it a performance advantage at the high end. AArch64 has adopted some of the CISC performance advantages and is using higher performance core designs but has so far been unsuccessful at dethroning x86-64 despite expensive attempts. Apple likely has the best chance but so far has chosen not to directly attack x86-64 strongholds.

 Status: Offline
Profile     Report this post  
fishy_fis 
Re: Why x86 doesn't need to die
Posted on 13-Apr-2024 15:53:55
#15 ]
Elite Member
Joined: 29-Mar-2004
Posts: 2163
From: Australia

@ppcamiga1

What the hell are you talking about you absolute fruit loop?

Amiga x5000 vs i9-13900k/Winuae for CPUBench both Running OS4.1 (while, running YouTube in the background, re-encoding a video and WinUAE set to low priority, so these scores are much lower than they can be):

p5040@ 2200 mhz x5000 040 / winuae i9 13900k
CPU max MIPS 4389 / 12563

RAM
READ32: 405 / 6627
READ64: 727 / 9260
WRITE32: 684 / 6469
WRITE64: 683 / 8953
WRITE: 2391(TRICKY) / 3424 (TRICKY)

VIDEO BUS
READ: 20 / 517
WRITE: 540 / 610

BOGOMIPS
calibrating delay loop..
ok - 352.00 BogoMips / 848 BogoMips

Relative to 10 Iterations and the 8191 Array Size:
High MIPS: 2315 / 2889
Low MIPS: 131.3 / 357

Results are much, much lower than using 68k emulation, but even with a very suboptimal setup on a machine 18 months old using inefficient ppc emulators its *still* much faster than any real ppc hardware available to AmigaOS.
As for ARM, absolute nonsense. Even an RPi5 is a heck of a lot more powerful than your G4, and thats definitely not expensive. And dont try the "but thats not a desktop" nonsense. It can run a heck of a lot more pci express video cards than your beloved G4, and utilize more of its potential too.

Are you too obsessed with your little crusade that it failed to register to you that many of us that you seem to think have a crusade against pp have ppc hardware and use OS4 and MOS as well?
Believe it or not, but wanting something other than a dead end for a hobby doesnt equate to being on a crusade against what it is that needs to be replaced.

Last edited by fishy_fis on 13-Apr-2024 at 04:01 PM.
Last edited by fishy_fis on 13-Apr-2024 at 03:56 PM.

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Why x86 doesn't need to die
Posted on 13-Apr-2024 20:30:18
#16 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12856
From: Norway

@fishy_fis

"much faster than any real ppc hardware available to AmigaOS."

thats a lie, FPU emulation is really bad.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
Karlos 
Re: Why x86 doesn't need to die
Posted on 14-Apr-2024 2:07:45
#17 ]
Elite Member
Joined: 24-Aug-2003
Posts: 4491
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@NutsAboutAmiga

That was true when the csppc emulation was first added. At least according to some tests I did at the time. I don't know how true it still is. We need a reproducible benchmark. I propose the following as it's very FPU intensive: https://mzucker.github.io/2016/08/03/miniray.html

_________________
Doing stupid things for fun...

 Status: Online!
Profile     Report this post  
ppcamiga1 
Re: Why x86 doesn't need to die
Posted on 14-Apr-2024 16:44:28
#18 ]
Cult Member
Joined: 23-Aug-2015
Posts: 829
From: Unknown

per core my i9 is only 3.5 faster than my g4 1.4 MHz

in turbo mode fans start and my laptop try to took off

I have to slowdown ppc emulator to run laptop in normal mode

in normal mode emulator is 5 times slower than my G4

my g4 is still almost 2 times faster than my rpi

even more than 68k emulator on my rpi

nothing compare to year 2002

when my cheap pc run 68k software 10 times faster than any real 68k

without any problems

and first peg run 15 times faster than any real 68k

after more than 20 years still no working real native amiga gui on x86/arm

so no reason to switch as in year 2000

Last edited by ppcamiga1 on 14-Apr-2024 at 04:45 PM.

 Status: Offline
Profile     Report this post  
Karlos 
Re: Why x86 doesn't need to die
Posted on 14-Apr-2024 17:12:29
#19 ]
Elite Member
Joined: 24-Aug-2003
Posts: 4491
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@ppcamiga1

Something seems to be wrong with your UAE machine or configuration if you're only getting 10x faster for 68K in UAE. I run WinUAE under wine on Linux and if I use pure 68040 JIT, no cycle emulations for memory etc, code runs way faster, even on an old 8th gen i7.

I can't speak for PPC emulation though. I definitely had issues with FPU performance on it back when I did try but that's some time ago now and we're on a new major version since then.

_________________
Doing stupid things for fun...

 Status: Online!
Profile     Report this post  
NutsAboutAmiga 
Re: Why x86 doesn't need to die
Posted on 14-Apr-2024 20:20:14
#20 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12856
From: Norway

@ppcamiga1

Laptops have tendency to overheat, and clock down.

And that was really true for old x86 cpu’s, the m1 and m2 cpu’s run really cold.

Last edited by NutsAboutAmiga on 14-Apr-2024 at 08:21 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
Goto page ( 1 | 2 Next Page )

[ home ][ about us ][ privacy ] [ forums ][ classifieds ] [ links ][ news archive ] [ link to us ][ user account ]
Copyright (C) 2000 - 2019 Amigaworld.net.
Amigaworld.net was originally founded by David Doyle