Click Here
home features news forums classifieds faqs links search
6055 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
44 crawler(s) on-line.
 10 guest(s) on-line.
 0 member(s) on-line.



You are an anonymous user.
Register Now!
 AASAAN:  7 mins ago
 Hammer:  9 mins ago
 lylehaze:  18 mins ago
 bhabbott:  25 mins ago
 MEGA_RJ_MICAL:  38 mins ago
 kolla:  43 mins ago
 THRIVECoworkingCumming:  1 hr 47 mins ago
 JKD:  1 hr 48 mins ago
 tygre:  3 hrs 46 mins ago
 dalek:  4 hrs ago

/  Forum Index
   /  Amiga General Chat
      /  Port AmigaOS 4 to x86
Register To Post

Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 Next Page )
PosterThread
amigang 
Re: Port AmigaOS 4 to x86
Posted on 6-May-2022 13:26:41
#21 ]
Super Member
Joined: 12-Jan-2005
Posts: 1850
From: Cheshire, England

I think AmigaOS should be ported to say a Virtual Processor, I remember reading about Tao Intent Technology was basically that.

But we kind of got that with UAE, emulating the 68k and limited PPC emulation. I kind thought why not just start emulating fake 68k hardware. We have Apollo and 68080 as a good bases, maybe that should be added to UAE. Why not emulate a multicore 68k.

I always had the thought why not make a version of AmigaOs thats more aware that its on emulated platform.

Rabbit Hole is handy little feature I like, i wish it could even more ingrate so that say for some apps it could map a host program full screen within a workbench window and mouse control is handled slightly better. I understand the difficulties of doing this but it would be cool if host program could be blended even more with OS.

then you could deploy AmigaOS on all platforms, and create a sort of Amiga Anywhere platform Just an idea

_________________
AmigaNG, YouTube, LeaveReality Studio

 Status: Offline
Profile     Report this post  
Karlos 
Re: Port AmigaOS 4 to x86
Posted on 6-May-2022 13:38:27
#22 ]
Elite Member
Joined: 24-Aug-2003
Posts: 2642
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@amigang

Quote:
I think AmigaOS should be ported to say a Virtual Processor, I remember reading about Tao Intent Technology was basically that.


I agree. My choice of virtual processor is a subset of the 68K architecture:

* 68020+ user mode integer operations and addressing modes.
* Optional 68040/60 user mode FPU for single and double precision operations.
* Optional additional support for extended precision on emulator targets with FPU.
* Optional additional support for 68882 transcendental operations on emulator targets with FPU.
* Additional hunk/elf metadata to assist AOT translation of binaries to host native code.

We already have highly mature JIT based execution for this and furthermore, the above subset should be executable on any 68020+ machine (which is a reasonable minimum compatibility in my view). I say the FPU is optional but in practise I think supporting the minimum 040/060 subset is a good idea for any new hardware solutions.

On the flipside, I'd rather make the OS less aware of what it is running on by improving the virtualisation around it and reducing the burden of hardware support on the 68K side.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
kolla 
Re: Port AmigaOS 4 to x86
Posted on 6-May-2022 14:41:07
#23 ]
Elite Member
Joined: 21-Aug-2003
Posts: 2123
From: Trondheim, Norway

@Karlos

Quote:
I'd rather see it backported to 68K


Geh no, please, keep OS4 on PPC, it’s fine there! Good riddance!

Let 68k OS move forward on its own, please. There’s enough “os4” bugs and nonsense in Os 3.2 already, no need for more of that, really!

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

 Status: Offline
Profile     Report this post  
kolla 
Re: Port AmigaOS 4 to x86
Posted on 6-May-2022 14:51:24
#24 ]
Elite Member
Joined: 21-Aug-2003
Posts: 2123
From: Trondheim, Norway

@Karlos

Quote:
On the flipside, I'd rather make the OS less aware of what it is running on by improving the virtualisation around it and reducing the burden of hardware support on the 68K side.


Then you’re essentially stuck to UAE, all though it already provides a lot of resources through 68k native uae specific resources/drivers (like uaehf.device etc).

But - I thought you said earlier that you want direct access to PCI bus etc Amithlon style? Now you seem to want the opposite.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

 Status: Offline
Profile     Report this post  
Karlos 
Re: Port AmigaOS 4 to x86
Posted on 6-May-2022 14:57:04
#25 ]
Elite Member
Joined: 24-Aug-2003
Posts: 2642
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@kolla

Quote:
But - I thought you said earlier that you want direct access to PCI bus etc Amithlon style? Now you seem to want the opposite.


No, the Amithlon model permits both. As I said elsewhere, out of the box gfx/audio/networking/usb via host virtualisation is a must have IMO, with the ability to provide direct access for hardware that may have existing PCI drivers the user may prefer is equally desirable. It's all about user choice and experience.

Just to reiterate, I don't mind if the OS is ported or not. What I care about is standardising around 68K as a Common Language Runtime ecosystem for Amiga application binaries that can be compiled once to run on classic, FPGA, OS4, MorphOS, UAE etc. The Amithlon idea is just an idea of what a "high end" user experience could be in that ecosystem.

Last edited by Karlos on 06-May-2022 at 03:07 PM.
Last edited by Karlos on 06-May-2022 at 03:02 PM.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Trixie 
Re: Port AmigaOS 4 to x86
Posted on 6-May-2022 15:08:43
#26 ]
Amiga Developer Team
Joined: 1-Sep-2003
Posts: 2058
From: Czech Republic

@kolla

Quote:
There’s enough “os4” bugs and nonsense in Os 3.2 already

Which are these, out of curiosity? I very much doubt Olsen and Thor would decide to backport something that is not up to scratch.

_________________
The Rear Window blog

AmigaOne X5000/020 @ 2GHz / 4GB RAM / Radeon RX 560 / ESI Juli@ / AmigaOS 4.1 Final Edition
SAM440ep-flex @ 667MHz / 1GB RAM / Radeon 9250 / AmigaOS 4.1 Final Edition

 Status: Offline
Profile     Report this post  
matthey 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 0:11:22
#27 ]
Super Member
Joined: 14-Mar-2007
Posts: 1502
From: Kansas

amigang Quote:

I think AmigaOS should be ported to say a Virtual Processor, I remember reading about Tao Intent Technology was basically that.


Emulating failure is the easiest kind of emulation.

amigang Quote:

But we kind of got that with UAE, emulating the 68k and limited PPC emulation. I kind thought why not just start emulating fake 68k hardware. We have Apollo and 68080 as a good bases, maybe that should be added to UAE. Why not emulate a multicore 68k.


Toni said, "No?"

Vampire emulation in WinUAE
http://eab.abime.net/showthread.php?t=84264

I will break down his arguments against emulating the Vampire.

Toni Wilen Quote:

Because it is just fast Amiga that is never going to be "finished" due to being reprogrammable, it is not interesting old piece of retro hardware.


This is partly true. Many modern high performance processors have upgradeable firmware/microcode and this is becoming more common in order to fix security problems and major bugs. I can see some processors continuing to move in the direction of blurring the lines with, for example, eFPGA blocks to offer the flexibility to customize a processor.

https://www.achronix.com/sites/default/files/docs/Speedcore_eFPGA_Product_BriefPB028.pdf

VLIW processors have "code morphing" which is part hardware and software but regular RISC/CISC processors may continue to develop flexible features for the CPU cores themselves. There are already several closely integrated CPU cores with a FPGA on the same die available. Using eFPGA blocks would be an ideal way to provide chipset simulations without the extra cost of a separate FPGA chip for a 68k game machine since we are talking about retro hardware.

Toni Wilen Quote:

Also all the extra instructions etc are in my opinion pointless that no one is going to use seriously anyway (except for demonstration purposes) and if someone does, it only causes fragmentation.


Again, partly true. There is minimal benefit in creating ISA changes without a large enough hardware/user base to become popular. Some acceleration from specialized code in drivers and codecs is likely possible and practical though.

Toni Wilen Quote:

It looks like Vampire has same problem that many (nearly all!) modern Amiga expansion have (that needs custom drivers): hardware is brilliant but then the design is force-fed to users and especially to developers without enough knowledge of AmigaOS internals and how hardware interface should work to be Amiga-like, result is weird address space, alien non-Amiga interface and not really compatible if custom drivers don't find the hardware and requiring to "upgrade" existing low level software to stay compatible. And still expect interest from developers when most of low level debugging utilities won't work and there is no way to do any quick testing with emulator first. (because hardware simply is too different than anything else).

...

If you are going to make Amiga compatible hardware, you need to work with AmigaOS, even if it it causes annoying restrictions. You shouldn't just bypass them with hacks because it is easy and AmigaOS limits annoy you.

If you have custom hardware, you should have also boot ROM that does the initialization. Thats how all other accelerators work, if they need HD controller drivers, add extra RAM etc.

And so on. Lack of proper firmware is IMHO the biggest problem with many "modern" expansions.


This is a valid point. Gunnar is responsible for hackish hardware and cutting corners in places.

Toni Wilen Quote:

But there is not enough developers, there is no compiler support, there is no debugging support. Everything needs to be done before actual support happens. It won't.


Partially true. There is partial developer tool support from the vasm assembler for example. There may be a debugger or disassembler around too. Still, there is no compiler support and unlikely to be any. I believe that part of the problem is an ISA which is way too complex with too many registers added everywhere. It should have been 16 integer 64 bit registers (perhaps configurable as 16 integer 32 bit data registers and 8 integer 64 bit address registers) but a large register file in a low clocked core doesn't have to worry as much about timing constraints. In other words, the ISA is optimized for high performance in a FPGA which doesn't make sense. I believe that part of the problem is that the developer community was not consulted about the ISA but rather had it forced on them instead of accepting it. I tried to bring in other respected developers and get input from other vbcc developers but that was seen as a threat to Gunnar's power. It reminds me of the Natami forum when Gunnar suggested adding a whole bunch of registers and the general consensus was that we didn't need them which made him very irritated and he ignored our input. I came up with a practical encoding for 16 FPU registers that he nixed because 16 wasn't enough. The complex bizarre FPGA optimized ISA by one man likely has a lot to do with why there is no support. ISA/ABI development, compiler development and OS development should be done with cooperation together by teams. I say that even as I was part of the Apollo Team but the "team" was really just a source of ideas for the "leader" who micromanages all the decisions.

 Status: Offline
Profile     Report this post  
Nonefornow 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 0:41:22
#28 ]
Regular Member
Joined: 29-Jul-2013
Posts: 307
From: Greater Los Angeles Area

@paolone

Quote:
You ask somebody to do something it's already been available for YEARS in the shape of AROS,


Very strong and valid point.

Like you I am perplexed by what appears to be a constant search for an alternative OS to run on a PC when AROS is already here.

Maybe it is time for all those conflicting parties to drop their esoteric projects and initiatives and pool their resources to advance AROS development.





 Status: Offline
Profile     Report this post  
BigD 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 1:24:37
#29 ]
Elite Member
Joined: 11-Aug-2005
Posts: 6414
From: UK

@Thread

If you want x86 you've got AROS, MorphOS will get there eventually but AmigaOS 4 will die with Hyperion it seems replaced with AmigaKit's System 54, whatever that is (said like Andy Crane being unimpressed with AGA's "millions of colours" brag)!

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

 Status: Offline
Profile     Report this post  
agami 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 2:24:55
#30 ]
Cult Member
Joined: 30-Jun-2008
Posts: 964
From: Melbourne, Australia

@Nonefornow

Quote:
Maybe it is time for all those conflicting parties to drop their esoteric projects and initiatives and pool their resources to advance AROS development.

AvengAROS, assemble!

Eh, it sounded good in my head.

_________________
All the way, with 68k

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 6:02:35
#31 ]
Elite Member
Joined: 29-Oct-2012
Posts: 2541
From: Germany

@ppcamiga1 Quote:

ppcamiga1 wrote:
Port AmigaOS 4 to x86 viable way.

There's no viable way for AmigaOS 4 and you, as THE die-hard fan, should know it.

OS4 is on a so bad shape that A-EON is trying to rewrite parts of it (or even all, due to the situation).

But even if it that wasn't true (but it's true), an OS4 port is unlikely to come because it looks like that Hyperion has no rights for some components, which were licensed only for PowerPC.
Plus, add the fact some core developers have repeatedly shown scorn against x86 ("the evil"?!?).

Summing all up, you don't need to be a fine logician to understand yourself (well, hard work with you, I know!) that you're asking for the impossible...
Quote:
Our community has already fast and cheap x86 solution for more than twenty years.
It is uae with JIT.
Nobody made new software for it.

Guess why: it's an emulator! Do you know the purpose of an emulator? You clearly show absolutely no understanding of the tool, here (but nothing new, eh!).
Quote:
It is simple.

Then... do it yourself, great genius!
Quote:
It is not as good as Windows/Linux/OSX.
Nobody will spend ten or more time to do exactly the some work
on exactly the some computer using any Amiga OS or it clone.
Any sane person will just switch to Windows/Linux/OSX.
Amiga OS or it clone on commodity hardware may be 10%
maybe 15% worse Windows/Linux/OSX.
No more.
It have to have all new techologies that Windows/Linux/OSX has.
It have to have drivers as good as Windows/Linux/OSX has.
It must be something like Amiga Os X.
Amiga GUI and graphics on unix base.
Amiga experience on modern base.
The only viable way to commodity hardware is
to made good Amiga Os gui and graphics open source clone.
No exec, no dos, no devices, no libraries
just Amiga gui and graphics on top of unix.

Other people already answered here: the solution is already available. Since YEARS.
Quote:
Who should do it ?

You! "It is simple." YOUR words!
Quote:
Attacks on Amiga Os 4 or ppc will give nothing.

Who has interesting on "attacking" (sic!) dead corpses?

You remind me a scene of famous movie. Santino was laying down, already dead, and people was still shooting at him...
Quote:
It is road to nowhere.

Be honest only ONCE in your life: this is what you, as THE OS4 die-hard fan, do NOT want to see. "Attacks" (ri-sic!) to your beloved OS4 (in PowerPC sauce, of course)...
Quote:
All that people that waste time on attacks on Amiga Os 4 or ppc

I revel you a secret: people is free to criticize everything, without that it means that this is an attack.

It's elementary logic, but it's something which should be highlighted in your (desperate) case.
Quote:
should work on zune etc.

Not needed: "It is simple", so you are the best candidate to work on it. Go on, and finally show something concrete!
Quote:
szulc, shonweiss, karlos, onetimer, di mauro, bison, neuf etc.

The legion of (your) enemies (at least what YOU, as THE OS4 die-hard fun, recognized as that)...

I reveal you another secret (it's the day, hey!): we've a real life and much more important things to do than shooting at cadavers.

Nobody of us contributed to the current OS4 status: you've to offer your congratulations to Hyperion, the developers which worked for it (most of them for free! Ah, the glory of working on "the named o.s."" doesn't deserve a payment...), and of course to its CEO which proved to be a better manager than Commodore's ones...
Quote:
How?

It need to be done on ppc for speed and compatibility reasons.

I reveal you another couple of secrets.

First, x86 machines are faster than PowerPC ones. So, if you look for speed, PowerPC is NOT the way to go.

Second, and usually, an emulator gives better compatibility compared to the dedicated machines. Any reference to the Tabor and its phenomenal FPU is not a coincidence. As well as WinUAE.
Quote:
First ppc aros should be fixed to add integration like on MOS or Amiga Os 4.
Then zune should be fixed to reach at least mui 3.8 compatibility.
Then good gui builder should be made for it.
Then zune should be made more modern.
After that we may switch to linux still on ppc.
Then port lo and decent webbrowser to it.
Also new blender and good developer editor should be done.
When it all be done We may use commodity hardware.

Of course zune should also work on 68k and ppc.

OK, since we're at the end, let's call them as they are: pure bullshits.
Quote:
Anybody who want to use it should be free to use it where want.

Like what the "Dynamic Duo" did with Timberwolf's sources?
Quote:
Cooperation is important to keep compatibility.

Then I want to see YOUR cooperation. But don't make me become decrepit...

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 6:19:00
#32 ]
Elite Member
Joined: 29-Oct-2012
Posts: 2541
From: Germany

@bison Quote:

bison wrote:
@ppcamiga1

Quote:
szulc, shonweiss, karlos, onetimer, di mauro, bison, neuf etc.

There's no interest on my part. I'm already working on two projects that are behind schedule.

Same for everybody here, besides the above genius, of course.

Thanks for quoting him and let me know that he cited me, because I stopped reading his frenzies long time ago and I've missed his recent "pearls" having me as one of the characters.


@hardwaretech Quote:

hardwaretech wrote:
Have you been failing the developments of AX-runtime on Aros Exec?
Sound like they already made progress on getting the bugs out of it.
Some basic apps have been included in the test app including Final Writer.
Since Windows now has a Linux subsystem some users are trying to get it to run under that.

It works since yesterday: a new Ax Runtime release should come soon.


@Karlos Quote:

Karlos wrote:
@michalsc

I'm pretty sure it's a troll post.

Well, he IS a troll. That's fully granted.


@michalsc Quote:

michalsc wrote:
@Karlos
Quote:
I'm pretty sure it's a troll post.


Of course it is a troll post. Still, I don't like misspelling my name on purpose.

Mispelling names is a clear indication of his poor and limited mind: it's the only way that the guy has to attack (which IS the case) what he recognized as his enemies, in order to satisfy his ego.

It's what in Italy we're used to say: "la consolazione degli stupidi" (rough translation: the consolation of stupid persons).

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 6:39:52
#33 ]
Elite Member
Joined: 29-Oct-2012
Posts: 2541
From: Germany

@agami Quote:

agami wrote:
Piggy-backing on top of other open source operating systems and kernels is about the only "life-boat" option remaining.

Also, just to be absolutely clear. Anybody who these days dumps on PPC Amiga streams, myself included, only does half as much as @ppcamiga1 dumps on 68k, AGA, Vampire FPGA, AROS x86 and UAE.

UAE is not a port of Amiga OS 3.x to x86. It is an application running on top of another operating system primarily intended to enjoy the past, so of course there would be very little interest in developing new software to run in such an environment.

Porting any Amiga OS, or its cousins, to x86 would naturally "die on the vine" if the port did not also bring new and improved development tools and modern sub-systems. If you don't make it easy for developers to bring their software to the platform, then very few will.

Porting AmigaOS 4 to x86 at this point is no longer worth the cost. Lots of work with minimal benefits. I look forward to seeing how far the MorphOS team has progressed with their x86 port, but of course I am not holding my breath.

Let's be realistic: Amiga and post-Amiga computers and/or oses are either retro or hobby projects.

There's no future neither any chance to revive it, in any shape.

And we're here only because we were mostly Amiga users, loved those machines, and still dream about them and that "golden age".

Since Amiga/o.s. died, we've chosen one of its post "incarnations", and we like to play with it. That's it.

It's just an hobby, like any other. Albeit for some it's quite a toxic one...

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 6:51:32
#34 ]
Elite Member
Joined: 29-Oct-2012
Posts: 2541
From: Germany

@Karlos Quote:

Karlos wrote:
@kolla

Quote:
But - I thought you said earlier that you want direct access to PCI bus etc Amithlon style? Now you seem to want the opposite.


No, the Amithlon model permits both. As I said elsewhere, out of the box gfx/audio/networking/usb via host virtualisation is a must have IMO, with the ability to provide direct access for hardware that may have existing PCI drivers the user may prefer is equally desirable. It's all about user choice and experience.

Just to reiterate, I don't mind if the OS is ported or not. What I care about is standardising around 68K as a Common Language Runtime ecosystem for Amiga application binaries that can be compiled once to run on classic, FPGA, OS4, MorphOS, UAE etc. The Amithlon idea is just an idea of what a "high end" user experience could be in that ecosystem.

That's more or less what I've proposed long time ago.

The best solution for running 68K/Amiga software is a "virtualizer": a tool which allows to execute 68K binaries "transparently", well integrated to the host system.

Something like VAMOS, but with a broader scope. I say VAMOS because a Python-based solution is much easier to implement, maintain, and enhance. And I prefer a pure Python tool, instead of integrating an external 68K core/emulator (like VAMOS do), for the same reasons (easier to expand / hack / do whatever).

Granted that it's quite evident that 68K is THE preferred hardware platform / ISA for Amiga aficionados. Not even counting the immense software library developed for it, which other hardware platforms lack (there's no competition. At all!).

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 7:50:04
#35 ]
Elite Member
Joined: 29-Oct-2012
Posts: 2541
From: Germany

@matthey Quote:

matthey wrote:
amigang Quote:

I think AmigaOS should be ported to say a Virtual Processor, I remember reading about Tao Intent Technology was basically that.

Emulating failure is the easiest kind of emulation.

Emulation has much less chances to fail compared to all other alternatives.

WinUAE clearly proven it.

But much better can be done with a "virtualizer", like I've said before.
Quote:
Quote:
Toni Wilen [quote]
Because it is just fast Amiga that is never going to be "finished" due to being reprogrammable, it is not interesting old piece of retro hardware.

This is partly true. Many modern high performance processors have upgradeable firmware/microcode and this is becoming more common in order to fix security problems and major bugs.

But you cannot fix frontend or backend issues. Even because microcode use is quote limited, and for good reasons: most of the executed instructions use no microcode, and that's the reason why they are so fast.
Quote:
I can see some processors continuing to move in the direction of blurring the lines with, for example, eFPGA blocks to offer the flexibility to customize a processor.

https://www.achronix.com/sites/default/files/docs/Speedcore_eFPGA_Product_BriefPB028.pdf

I don't see the difference, in this context, with other FPGAs.
Quote:
VLIW processors have "code morphing" which is part hardware and software

This is only for some VLIW processors. Specifically, for Transmeta and nVidia ones.
Quote:
but regular RISC/CISC processors

There are essentially no RISC processors. Since very long time only CISC processors were produced and that's what we're using practically everywhere.

What many companies and academics (sic!) call and sell as RISCs aren't RISCs, rather CISCs. This is the RISC propaganda which tried (and still continue) to hide the complete failure of the RISC paradigm.
Quote:
may continue to develop flexible features for the CPU cores themselves. There are already several closely integrated CPU cores with a FPGA on the same die available. Using eFPGA blocks would be an ideal way to provide chipset simulations without the extra cost of a separate FPGA chip for a 68k game machine since we are talking about retro hardware.

How? It's not clear from the document that you provided.
Quote:
Quote:
Toni Wilen [quote]
But there is not enough developers, there is no compiler support, there is no debugging support. Everything needs to be done before actual support happens. It won't.

Partially true. There is partial developer tool support from the vasm assembler for example. There may be a debugger or disassembler around too. Still, there is no compiler support and unlikely to be any. I believe that part of the problem is an ISA which is way too complex with too many registers added everywhere.

No. The clear target for Vampire is assembly programming (and direct hardware hitting).

There's no doubt about it, according to the several times that it was repeated by Gunnar or some other Vampire member.

For this reason, an assembler (and disassembler) is enough: no high-level compilers support is required.

And for the same reason the Vampire will be limited to a niche. Which isn't unexpected, BTW...
Quote:
It should have been 16 integer 64 bit registers (perhaps configurable as 16 integer 32 bit data registers and 8 integer 64 bit address registers) but a large register file in a low clocked core doesn't have to worry as much about timing constraints. In other words, the ISA is optimized for high performance in a FPGA which doesn't make sense. I believe that part of the problem is that the developer community was not consulted about the ISA but rather had it forced on them instead of accepting it. I tried to bring in other respected developers and get input from other vbcc developers but that was seen as a threat to Gunnar's power. It reminds me of the Natami forum when Gunnar suggested adding a whole bunch of registers and the general consensus was that we didn't need them which made him very irritated and he ignored our input. I came up with a practical encoding for 16 FPU registers that he nixed because 16 wasn't enough. The complex bizarre FPGA optimized ISA by one man likely has a lot to do with why there is no support. ISA/ABI development, compiler development and OS development should be done with cooperation together by teams. I say that even as I was part of the Apollo Team but the "team" was really just a source of ideas for the "leader" who micromanages all the decisions.

The thing is very simple: Gunnar likes to decide everything, because the project is essentially his own. He has some ideas about what Vampire should be, and he's pursuing his direction.

And I've nothing to say about it, because I would have done the same if I had his expertise and chance. And you know why? Because, at the very end, he was able to deliver! Deliver a product to people who like exactly what he proposed to them.

When I was at Intel I've learned a simple but quite effective mantra which is circulating on the company: deliver is the king! And with this mindset we worked. And delivered. And achieved success. And gained market.

I think that, as computer architecture passionate people, we recognize mistakes and bad design decisions which Gunnar made with the 68080 (and SAGA too, if we want to extend a bit the topic).

But think about it: what could have happened if you were the 68080 designer and you delivered the product according to your vision? Do you think that nobody could have criticized it / your work?

How many times have we (me, you, megol, Gunnar, etc.) discussed about ISA (and chipset) topics? How many times we had different ideas and disagreed? Countless...

Because this is the reality: everyone has HIS vision about how a certain ISA should be designed and implemented. And it's VERY DIFFICULT to find a common agreement. Impossible, I would say, after so much discussions.

At least Gunnar was able to deliver. Kudos to him!

For this reason I think that we should separate the two things, and don't mix them.
One thing is our evaluation about ISA, chipset, and whatever technical topic.
Another thing is the products on the market.

To be more clear, finger pointing Gunnar for what he did doesn't make sense: those are pointless complaints (here I also see a little bit of envy, eh! I also envy him because he produced something which I wasn't able to do).
So, let's limit our analysis to the technologies which were delivered and/or about how personal vision of how things should be done.

 Status: Offline
Profile     Report this post  
Karlos 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 14:58:51
#36 ]
Elite Member
Joined: 24-Aug-2003
Posts: 2642
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@cdimauro

Quote:
That's more or less what I've proposed long time ago.


"As terrible as Zoidberg's idea is, it's brilliant when I have it!" -- Bender Bending Rodriguez

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 20:56:00
#37 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12197
From: Norway

@Trixie

From what I have read on Amiga.org, some places there where some small changes to some of the commands, not sure what, I don’t remember, no one compiled a list of this problems Kolla found. And not sure everyone will agree that are actual bugs.

There absolute difference in how AmigaOS4.x behaves vs AmigaOS3.x, for example AmigaOS4.x default to true color chunky modes, if no mode id is set in OpenScreenTags, that one is problematic, as a lot programs, and games are hard coded, to 40 bytes per row, planar screens as default, AmigaOS4.x uses 80 bytes per row (640 pixel on lowres), stuff like that does not make things easier.

Raw key codes, is also problematic, Amiga keyboard and AmigaOS4.1 usb keyboards, do not use the same key codes. That’s problematic, when your running old 68k program. it be good ide if input.device know if the program is 68K program or PPC program, maybe Exchange app can fix it, yet it is a problem.

The EClock value is different.

Last edited by NutsAboutAmiga on 07-May-2022 at 09:17 PM.
Last edited by NutsAboutAmiga on 07-May-2022 at 09:03 PM.
Last edited by NutsAboutAmiga on 07-May-2022 at 09:00 PM.

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

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 22:18:48
#38 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12197
From: Norway

@Karlos

I think the problem with that idea, is that you trying to force, MC680x0 code into becoming something it was never was intended for, what is the standard for MC680x0 Amiga assembler code exactly.
Some time you bang the hardware, sometime you can’t, sometime you take over the OS, and sometime you can’t, while MC680x0 is shared between MOS PPC, AOS PPC, AROS 68K, AOS 68K, Amitlon the ABI is somewhat not standardized.

(Naturally native ABI is faster, not sure all new stuff is available on the EMU 68K ABI, or if that only in native ABI of AmigaOS4.x ppc.)

Tools made to make WHDLOAD games, and ADF’s games with boot loaders, people continue to write OCS/AGA demos, the adoption of RTG/P96/CGX is slow. I think some action from 68K community as defiance agents the NG systems. The other part is old tools to make games and demos, that’s not up to date. Its not the OS that problem its community and slow adaptation of RTG, and lack of developers to maintain and update the software. Even worse who do you fix a 20-year-old program, there is no source code for, who be interested in disassemble and writing patches, rewriting or replacing old tools? What language will choice. Do you think 68k assembler or C/C++?

Last edited by NutsAboutAmiga on 07-May-2022 at 10:20 PM.

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

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 22:28:24
#39 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12197
From: Norway

@Karlos

I wonder if not Web assembly or LLVM or something like be better,
Emulating of really low-level stuff like flags, or interrupts, mmu’s and FPU, does it make sense? Wont be better to abstract that, let middle ware take of that.

I think creating our own programming language, that takes care of endianness can be a why to approach this, this why write this or that why, know what format its written, it be easy to know when you need to convert something, and if you have be endian CPU, you ca use best opcode to handle it.
I think its shameful that was not already done in C/C++.

What a sec, Hollywood programming language does this more or less, at least has become a why to make portable Amiga programs.

Last edited by NutsAboutAmiga on 07-May-2022 at 10:32 PM.

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

 Status: Offline
Profile     Report this post  
Karlos 
Re: Port AmigaOS 4 to x86
Posted on 7-May-2022 22:49:54
#40 ]
Elite Member
Joined: 24-Aug-2003
Posts: 2642
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@NutsAboutAmiga

Quote:
I think the problem with that idea, is that you trying to force, MC680x0 code into becoming something it was never was intended for.


In what sense? As with all binary object code the intention is that it is executed on something. Usually that's hardware processor, but there's no reason it has to be. No reason at all.

On Endianness...
Quote:
I think its shameful that was not already done in C/C++.


Why? What business is it of C or C++ how a given platform organises it's bits and bytes? The only times endianness is ever a factor are when:

1) You decide to access the contents of the same memory location using different words sizes (e.g. pointers to different integer types or messing around with unions in a risky way).
2) You are loading raw binary data in a fixed endian representation that is possibly different than your own.

Last edited by Karlos on 07-May-2022 at 10:55 PM.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 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