@ppcamiga1
Haven't you understood by now that Arm is also big endian?

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@MagicSN

ppcamiga is our favourite PPC troll. The chances are excellent he doesn't have any Amiga hardware or any kind.

_________________
Doing stupid things for fun...

@pixie

He doesn't care. If it's not PPC he just regurgitates the same bollocks.

_________________
Doing stupid things for fun...

@ppcamiga1

Quote:

ppcamiga1 wrote: treat pistorm as it is. pistorm is hardware that change amiga into mouse, joystick, and keyboard interface for rpi.

PiStorm is fundamentally just a faster 68k CPU with RAM. If its other assets such as RTG and SD Card are used then it gets closer to what you describe, but so does an accelerator card like the GVP G-Force 040 (released in 1992), which with the optional EGS 110/24 graphics card had RTG, SCSI2 and even serial and parallel ports, would change the Amiga even more into a 'mouse, joystick, and keyboard interface' for the GVP board. Yet nobody thought of it as such.

The PiStorm has a blisteringly fast 68k CPU which allows it to do stuff no classic Amiga could. But an actual 68k CPU made with modern technology could do the same (if anyone had the money to make one), so the PiStorm is really just a cheap way to get there. Most Amigas were designed to take accelerator cards so the PiStorm doesn't break the retro spirit.

An Amiga with PiStorm still runs the same OS, still has the same graphics and sound chipset, still runs the same programs without any modifications. For those of us who want to continue using classic machines in the retro spirit and also experience what it could have been like if 68k was developed further the PiStorm is a good choice.

OTOH the ARM CPU in the RPi could be used to run a more modern version of Amiga OS natively for even better performance. RPi 3 can run big-endian if you want (though I think it might be better to switch to little-endian since it is more common). That is what Commodore planned to do to the Amiga with Hombre, would would have had a 68k CPU and AGA chipset in a SoC for backwards compatibility.

@bhabbott

May I refer you to my earlier post. You could produce a physical 68100, completely ASIC, 3GHz, 100% 68020/68882 instruction set compatible with the extras from 040/060, pair it with modern RAM, PCIe graphics board and other expansions and he'd still go on about PPC. He's a troll. It's what he does.

_________________
Doing stupid things for fun...

@MagicSN

Quote:

MagicSN wrote: My personal expectation is you can only get the emulation crowd if you actually get GL Renderer. Wazp3D on software is too slow (tried this on QEmu), and why should someone buy a game he can only run in Software rendering when on the same machine he can run the Windows version in 3D Hardware ?

Because he's a masochist?

I ran Tomb Raider on my A1200 and my Vampired A600 via PC Task, and it was amazing to see my favorite 3D game on the Amiga. But on the A600 it only runs ~5fps which is barely playable, and 0.5fps on the A1200 which is a literal slideshow.

I would love to see Tomb Raider ported to the Amiga, but I know it won't happen. I don't have the energy to do it myself and I doubt anyone else does either. Furthermore I want to play the latest user-created Tomb Raider games that only work on a Windows PC. Unfortunately a mouse piddled on the graphics card and killed the AGP port on my Windows PC, so I had to fall back to the onboard graphics which is too slow.

If the goal is just to play a game you use the appropriate hardware for the job. But retro computing is more than that. I might stick to playing games on the Amiga now because I am sick of my 'modern' PCs breaking down all the time. Two others broke recently. One is totally dead and other one's power supply blew up with a bang! (can't replace it because the case is a non-standard size). The 386-SX motherboard I recently bought to make a retro PC also died. PCs suck!

@Karlos

Quote:

Karlos wrote: @bhabbott May I refer you to my earlier post. You could produce a physical 68100, completely ASIC, 3GHz, 100% 68020/68882 instruction set compatible with the extras from 040/060, pair it with modern RAM, PCIe graphics board and other expansions and he'd still go on about PPC. He's a troll. It's what he does.

Except that's not going to happen.

ppcamiga1 seems to be upset about 68k getting more love than ppc these days, but he also talks about sparc, mipc, and even itanium being OK. It's just 68k he's upset about, and the reason isn't the CPU so what is it? I'm guessing it's that 68k runs OS 3 which is getting real updates while OS4 isn't. In fact OS3.2 could be considered more advanced than OS4, suggesting it is the future of the Amiga. Classic Amiga owners like that development, while those like ppcamiga1 who hitched their wagon to NG Amigas don't. But it's not our fault if it didn't achieve what they were expecting. The failure of ppc is all on them.

Last edited by bhabbott on 06-Dec-2023 at 06:54 PM.

@bhabbott

Quote:

RPi 3 can run big-endian if you want (though I think it might be better to switch to little-endian since it is more common).

But in big endian you could run both systems, like in AmigaOS 4 or MorphOS

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@bhabbott

If I remember right the Tomb Raider "Source Code" which was released was some sort of ASM Source readout. So basically it would not work if you do not have an x86 (lots of x86 ASM still in there).

>If the goal is just to play a game you use the appropriate hardware for the job. But retro computing >is more than that. I might stick to playing games on the Amiga now because I am sick of my >'modern' PCs breaking down all the time. Two others broke recently. One is totally dead and other >one's power supply blew up with a bang! (can't replace it because the case is a non-standard size). >The 386-SX motherboard I recently bought to make a retro PC also died. PCs suck!

Absolutely agreed. When I first was planning the release of H2 (which is to be considered retro computing these days) the plans were along a digital release only.

I was getting MASSIVE amount of emails (Users, dealers, whoever,...) about "DONT DO THAT". People wanted a proper "box" like "back then". It is the retro spirit (it was not easy to find a producer for that).

In the end I went with that. The final box production (I got a trial version already) is probably MUCH MUCH superior to any PC Game box, and definitely looks better than the box of the WarpOS OS 3.x version. I also got the point of we need a printed manual (originally not planned).

Material of the box is a bit like of a boardgame box. When I put it besides a box of the Windows version - what a difference !!!

I got it now. If doing a game for AmigaOS these days you need to embrace the retro spirit.

@bhabbott

Of course it's going to happen! Matthey has been saying it for years now. Unless you are suggesting he's just full of warm gas...

But of course I jest. Perhaps under some controlled regression therapy, ppcamiga could show us where the bad 68K touched him, and we'd understand his... unique perspective.

And then when we've all recovered our composure, we can go on to talk about how we might look to bridge the gap between Emu68 and whatever arm binary blobs pass as vendor supplied drivers for the onboard hardware.

Last edited by Karlos on 06-Dec-2023 at 07:46 PM.

_________________
Doing stupid things for fun...

@Banging hardware..

You can actually peek poke CIAA/CIAB and some of CUSTOM registers,
if you have Nalle Puh running in the background.

Actually, I love if people messed around with NallePuh a bit,
we perhaps find a few bugs or perhaps someone has some idea about makeing it better.

Before you use CIAA/CIAB timer its common sense to check if timer is available,
using the ciaa.resouce and ciab.resource. You can’t do everything using resources, you need peek poke the counter value to hardware registers, someone quoted some thing about never writing direct to CIAA/CIAB chips.

When you read the ICR / CR registers, you reset the registers, this why should use the resources, but all other registers you can read from, or write to as long as you’re the owner timer.

I have made it so that CIAA/B interrupt rutine can be 680x0 or PowerPC code, so you can write in C/C++ code, and compile it with any C compiler if you like.

(Of course, you should not bang hardware, in particular when the hardware is emulated, it’s not optimal, naturally.)

Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:43 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:09 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:07 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:05 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 08:41 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 08:37 PM.

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

@Self-modifying code.

You can write self-modifying code, how do you think a JIT compiler work?
the thing about CPU with large instruction cache, is that it had old version of instructions in the cache, so you need to flush the cache before you execute it. this work nicely for any PowerPC code and used it to optimize a unroll routine to CPU cache and to leverage maximum registers, to improve out of order execution. (Sadly, my code only worked specific GCC version, I think I made some mistake about ABI.)

For 680x0 code its more complicated, but what I wrote about is also true there, but instruction cache is smaller, so you might not run into trouble, AmigaOS4.x has two modes to deal with 680x0 code, it JIT it and convert into PowerPC or to interparent it, this require the 680x0 program your trying to run, being added to compatibility preferences. This has to benefit, it makes more compatible with self-modifying code, and it also makes it easier to debug, (as when your executing JIT cache, it only knows the start program counter value before the JIT block, so debug output is imprecise, if your looking for a bug.)

E-UAE debugger is also actually advanced as well, this is also a good option for debuing 680x0 code.

Last edited by NutsAboutAmiga on 06-Dec-2023 at 08:59 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 08:56 PM.

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

@NutsAboutAmiga

Self-modifying code is usually only used to describe static code that patches itself at runtime by overwriting parts of itself with new code. A JIT doesn't do that, it generates code and puts it into some buffer for execution. It may well modify that code, but it doesn't modify itself. It has more in common with a loader in that respect.

_________________
Doing stupid things for fun...

@Karlos

Yeh, you run out of memory quite quickly if you can’t reuse memory… LOL…

To generate PowerPC code into memory block you need to allocate it as executable memory, or else the exception handler will be triggered.

modifying the exe files memory, can be done, but protected by default, so you need to unprotect it, I wont advice it, as you lose some of the memory protection, that cause stuff be lot worse to debug.

You don’t need to unpack code, we lots of memory, and SSD and M2 disks really fast, its not like space problem, its not like we are running web applications all day long. You need a good reason to do it.

Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:46 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:45 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:33 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:30 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:30 PM.
Last edited by NutsAboutAmiga on 06-Dec-2023 at 09:25 PM.

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

@NutsAboutAmiga

In my experience, self-modifying code is rarely done to save space. Instead it's used to concretise certain things that might otherwise be dependent on constantly evaluating something that's not going to change at runtime, such as code that depends on which CPU you are using that has a version for each.

_________________
Doing stupid things for fun...

@Karlos

More like binary patch? Wont it be tricky with different length of instruction on 680x0, to just overwrite a part.

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

@MagicSN

Quote:

MagicSN wrote: @Karlos Still right now PowerPC Amigas are the most powerful Amigas (especially x1000/x5000 systems).

I love your passion @MagicSN, but this statement doesn't really mean anything. As in, it has no value.
Right now, as we close out 2023, it would be like stating that some device is:
- the most powerful analogue modem, or
- the most powerful DVR, or
- the most powerful portable SACD player.

When we live in a world filled with:
- affordable 5G/4G cellular network connectivity
- movie and TV show streaming services, and
- music streaming services with all the music.


Quote:

at least for the stuff I am doing - there is not really an alternative to PowerPC Amigas.

This is really all you ever had to say.
It's OK to prefer something to which most other people don't have access.

I'm sure there's someone out there who can equally state that for at least what they're doing, there's no real alternative to:
- A restored Aston Martin DB5, or
- Microsoft's HoloLens 2, or
- A Steinway & Sons grand piano.

You (in this case I am actually referring to you @MagicSN) chose the pursuit on which you wish to spend your time, and then you chose the best tool for that pursuit that you could afford. But that doesn't mean that the tools you chose are the best for most people who have a varying interest in the same pursuit.

So when we consider these things, we can see that the attacks on PPC hardware, real or perceived, are not really senseless at all.

Last edited by agami on 07-Dec-2023 at 01:44 AM.

_________________
All the way, with 68k

@MagicSN

Quote:

MagicSN wrote: @Karlos Still right now PowerPC Amigas are the most powerful Amigas (especially x1000/x5000 systems).

I love your passion @MagicSN, but this statement doesn't really mean anything. As in, it has no value.
Right now, as we close out 2023, it would be like stating that some device is:
- the most powerful analogue modem, or
- the most powerful DVR, or
- the most powerful portable SACD player.

When we live in a world filled with:
- affordable 5G/4G cellular network connectivity
- movie and TV show streaming services, and
- music streaming services with all the music.


Quote:

at least for the stuff I am doing - there is not really an alternative to PowerPC Amigas.

This is really all you ever had to say.
It's OK to prefer something to which most other people don't have access.

I'm sure there's someone out there who can equally state that for at least what they're doing, there's no real alternative to:
- A restored Aston Martin DB5, or
- Microsoft's HoloLens 2, or
- A Steinway & Sons grand piano.

You (in this case I am actually referring to you @MagicSN) chose the pursuit on which you wish to spend your time, and then you chose the best tool for that pursuit that you could afford. But that doesn't mean that the tools you chose are the best for most people who have a varying interest in the same pursuit.

So when we consider these things, we can see that the attacks on PPC hardware, real or perceived, are not really senseless at all.

_________________
All the way, with 68k

@pixie

Quote:

pixie wrote: @cdimauro Didn't AROS had a proof of concept where multiple instances of it were running on linux? Perhaps with a bit of arexx glue they could be seen by the user as one system.

No exactly. AxRuntime allows to execute multiple AROS applications in parallel, as "sandboxed processes" on a Linux distro (which was properly setup).

It's not a PoC: it already works! I've also tested it on my Windows machine with the Linux subsystem (WSL2) and I was able to execute GUI applications.

deadwood also worked to have ScalOS running as desktop / Workbench replacement, in order to have a fully-fledged Amiga environment, but fixing it make it work on AROS requires still some work (however it looks very promising).


@MagicSN

Quote:

MagicSN wrote: @V8: >Under no circumstances whatsoever would I or any sane people provide my full name and address >and some legalese agreements that will then be handed over to Ben Hermans. >Please understand this. Anyone providing their name and address and some legal promises to Ben >Hermans will dearly regret it. >Do not do this. I would more say no sane developer would send unreleased commercial software to someone not under NDA. Where he does not even know the address. >Why do you even need this stuff? There is zero commercial value in this IP. I disagree. I take it is an ideological problem, but if it is I cannot help it. I won't send out Beta versions to un-NDA'ed people where I do not even know the physical address. And myselves I lack the legalese knowledge to set up a NDA. Nobody forces you to become Betatester. It is your choice. And it is my choice that I require an NDA for Betatester. I have around a dozen different testers, NDAs were never an issue.

Seriously? A NDA for testing a port of a videogame which was released DECADES ago for other platforms?

If this is really the case I expect only a single line there: "Don't distribute the binaries and the test results". Which is the only thing which COULD (!) have made sense...


@pixie

Quote:

pixie wrote: @ppcamiga1 Haven't you understood by now that Arm is also big endian?

So pixie, it's you that don't understand. Let me fix your question:

"Haven't you understood?"

Now I can answer you: yes!


@bhabbott

Quote:

bhabbott wrote: If the goal is just to play a game you use the appropriate hardware for the job. But retro computing is more than that. I might stick to playing games on the Amiga now because I am sick of my 'modern' PCs breaking down all the time. Two others broke recently. One is totally dead and other one's power supply blew up with a bang! (can't replace it because the case is a non-standard size). The 386-SX motherboard I recently bought to make a retro PC also died. PCs suck!

As I've previously stated, my Amiga 1200 died (video section not working anymore) after not even 4 years (in 1996): Commodore sucked A LOT!!!


@bhabbott

Quote:

bhabbott wrote: @Karlos Quote: Karlos wrote: @bhabbott May I refer you to my earlier post. You could produce a physical 68100, completely ASIC, 3GHz, 100% 68020/68882 instruction set compatible with the extras from 040/060, pair it with modern RAM, PCIe graphics board and other expansions and he'd still go on about PPC. He's a troll. It's what he does. Except that's not going to happen. ppcamiga1 seems to be upset about 68k getting more love than ppc these days, but he also talks about sparc, mipc, and even itanium being OK. It's just 68k he's upset about, and the reason isn't the CPU so what is it? I'm guessing it's that 68k runs OS 3 which is getting real updates while OS4 isn't. In fact OS3.2 could be considered more advanced than OS4, suggesting it is the future of the Amiga. Classic Amiga owners like that development, while those like ppcamiga1 who hitched their wagon to NG Amigas don't. But it's not our fault if it didn't achieve what they were expecting. The failure of ppc is all on them.

Still discussing about a troll which is REPEATING LIKE A PARROT always the same things with the solely purpose of defending his non-existent platform (there were never been any PowerPC Amiga)?


@NutsAboutAmiga

Quote:

NutsAboutAmiga wrote: @Self-modifying code. You can write self-modifying code, how do you think a JIT compiler work? the thing about CPU with large instruction cache, is that it had old version of instructions in the cache, so you need to flush the cache before you execute it. this work nicely for any PowerPC code and used it to optimize a unroll routine to CPU cache and to leverage maximum registers, to improve out of order execution. (Sadly, my code only worked specific GCC version, I think I made some mistake about ABI.) For 680x0 code its more complicated, but what I wrote about is also true there, but instruction cache is smaller, so you might not run into trouble, AmigaOS4.x has two modes to deal with 680x0 code, it JIT it and convert into PowerPC or to interparent it, this require the 680x0 program your trying to run, being added to compatibility preferences. This has to benefit, it makes more compatible with self-modifying code, and it also makes it easier to debug, (as when your executing JIT cache, it only knows the start program counter value before the JIT block, so debug output is imprecise, if your looking for a bug.) E-UAE debugger is also actually advanced as well, this is also a good option for debuing 680x0 code.

I really don't see the problem: just STOP using the JIT when the emulated 68k machine has disabled the cache. So, switching back to full emulation (interpretative mode).

OR clear the JIT buffers if the cache is enabled and the 68k code has cleared it (because it has made some changes on the existing code).

Simple, no?


@Karlos

Quote:

Karlos wrote: @NutsAboutAmiga Self-modifying code is usually only used to describe static code that patches itself at runtime by overwriting parts of itself with new code. A JIT doesn't do that, it generates code and puts it into some buffer for execution. It may well modify that code, but it doesn't modify itself. It has more in common with a loader in that respect.

Formally an application can also generate the code which it then can execute.

The one which was on my games had recurring patterns. Now I don't recall if I decided to generate it automatically in the final version, or just kept the full list as it is.
Quote:

Karlos wrote: @NutsAboutAmiga In my experience, self-modifying code is rarely done to save space. Instead it's used to concretise certain things that might otherwise be dependent on constantly evaluating something that's not going to change at runtime, such as code that depends on which CPU you are using that has a version for each.

It could also be used in different ways...


@agami

Quote:

agami wrote: @MagicSN Quote: MagicSN wrote: @Karlos Still right now PowerPC Amigas are the most powerful Amigas (especially x1000/x5000 systems). I love your passion @MagicSN, but this statement doesn't really mean anything. As in, it has no value.

No, it simply has no sense, at all.

As I've already stated several times, there were NO PowerPC Amigas. By definition!

As it can be clearly read from Commodore's official documentation, an Amiga is made of a chipset like OCS, ECS (and AGA, which is backward-compatible with them) AND a CPU from the Motorola 68000 family. Full dot!

After so long, is it that difficult t just open a FUNNY book, read it, and understand this elementary concept?!?

@agami

If you are not believing Amiga still has some value why are you even here ?

And my statement on "PowerPC Amigas right now are the most powerful Amigas right now" is based on actual measurable data. Frame rate in this case.

Might be PiStorm and QEmu also come there in 1-2 years, but right now that's the fact.

@cdimaura

>Seriously? A NDA for testing a port of a videogame which was released DECADES ago for other >platforms?

Don't play at being shocked. People want to buy these games. That YOU don't want to buy it does not mean others do not want. And I am not taking a risk here.

>As I've already stated several times, there were NO PowerPC Amigas. By definition!

There is one here right now. It has a PowerPC, it runs AmigaOS natively. It is an Amiga. BY DEFINITION.

>As it can be clearly read from Commodore's official documentation, an Amiga is made of a >chipset like OCS, ECS (and AGA, which is backward-compatible with them) AND a CPU from the >Motorola 68000 family. Full dot!

That is what we call a "Classic Amiga". Not all Amigas are Classic Amigas.

Stop trolling please. We waste each other's time.

Steffen

Last edited by MagicSN on 07-Dec-2023 at 07:03 AM.