ARM for the future?

[the_leander]

Both share the same endian design (as does the ARM, and NOT the X86).

As far as I know ARM is bi endian. Same goes for PPC, at least in some models.

[the_leander]

x86 is fine for performance, and it's true that it's come a long way since the 8086, it's still ugly as hell under the hood - the "four sort-of-but-not-really general-purpose registers" approach was obnoxious back in the day, and it's downright inexcusable now, with not only modern architectures but even the original 68000 providing a full complement of interchangeable registers. And it hasn't improved, either - they've just kept expanding the register width. Granted, a compiler can work out a good approach for shuffling values into and out of the registers, but for those of us who like assembler, it's pure aggravation.

1994 was just calling, they want their technical info back.

From wikipedia:

With the advent of the 32-bit 80386 processor, the 16-bit general-purpose registers, base registers, index registers, instruction pointer, and FLAGS register, but not the segment registers, were expanded to 32 bits. This is represented by prefixing an "E" (for Extended) to the register names in x86 assembly language. Thus, the AX register corresponds to the lowest 16 bits of the new 32-bit EAX register, SI corresponds to the lowest 16 bits of ESI, and so on. The general-purpose registers, base registers, and index registers can all be used as the base in addressing modes, and all of those registers except for the stack pointer can be used as the index in addressing modes.

Two new segment registers (FS and GS) were added. With a greater number of registers, instructions and operands, the machine code format was expanded. To provide backward compatibility, segments with executable code can be marked as containing either 16-bit or 32-bit instructions. Special prefixes allow inclusion of 32-bit instructions in a 16-bit segment or vice versa.

With the 80486 a floating-point processing unit (FPU) was added, with eight 80-bit wide registers.[17]

With the Pentium II, eight 64-bit MMX integer registers were added (MMX0..MMX7, which share lower bits with the 80-bit-wide FPU stack (st(0)..st(7))).[17] With the Pentium III, a 32-bit Streaming SIMD Extension (SSE) control/status register (MXCSR) and eight 128-bit SSE floating point registers (XMM0..XMM7) were added.

And that again increased with the advent of X64 (AMD Athlon64 onwards).

[the_leander]

Now that's funny!
RISC was supposed to be big.

In a sense they became what everyone suspected they would since at the heart of any modern X86/x86-64 processor is a risc core hooked up to some logic that provides compatibility with the x86 instruction set.

[the_leander]

Okay, my mistake. Are they really general-purpose, though? Looks from that excerpt like it's more just the gradual integration of specialized coprocessor hardware into the CPU than actual additional registers.

Well the excerpt is from here although I imagine there are much more technical resources out there or linked to within the article that would be better suited to answering your questions.

I'll still take either the 68k or ARM for assembler programming, in any case.

What about MIPS?

[the_leander]

care to explain why should anyone write their OS completely in asm code in a world of multicore GHz CPUs and storage that sizes in TBs.... instead of putting up various pictures?

The challenge of doing so. Menuette is very cool as projects go.  

But beyond that just because you have multicore cpus doesn't mean you shouldn't or needn't bother with optimising code. The major growth area in terms of development is in portable computing, starting with smart phones and continuing with tablets and netbooks. Heavy, unoptimised programs don't just mean poor performance, but decimation of battery life.

[the_leander]

I agree partially, but one should also note that today's compiler's do a pretty fine job of optimizing output asm code. Probably better then 99% of programmers who would try to replace them with a manually written asm/machine code. If one writes a elegant, properly structured program in a higher-level language such as C or JAVA, one will have a responsive program... Simple as that. No need to use asm today.

Tell me, when the assembly programmers have finished with you, where do you want me to send the flowers?

There are plenty of instances where being able to use asm is not only handy, but very necessary. Many small industrial micro controllers for instance.

[the_leander]

I haven't said I'd get rid of them completely

No you stated that it was completely irrelevant due to today's computing power.

Just that it is not needed to code a OS or applications 100% in assembler.

I don't think anyone said it is necessary, only that it is an interesting project to see just what it can be made to do as far as choice of computing language, as Karlos has said many times, pick the right language for the right job.

The reason you got the response you did was because what you wrote read like an excuse to code poorly as well as a complete blanket dismissal of asm.

[the_leander]

:roflmao:
 
Exactly how I feel about it, always been a big fan of RISC myself, would love to see a NG Amiga OS on a RISC based system.

So why discount x86? The heart of which is a RISC core and has been since around the time of the Cyrix 5x86. As far as I'm aware all currently produced x86 and x86-64 processors take this approach.

[the_leander]

I don't get this trashing of x86 that usually happens on amiga forums?

From the older set, by that I mean those of us who were using Amigas during their heyday it is somewhat understandable - Microsoft quickly replaced Atari as "the enemy". Watching technically superior hardware and software overrun by the PC industry with equipment that was "good enough" but quickly evolving was hard to do.

For those who came after C='s demise, especially with those who only showed up around the very late 90's onwards, I must confess I'm a little confused myself. Perhaps some of the zeal is down to wanting to be outside the mainstream, so as to gain a sense of superiority from those using commodity hardware. Tbh I suspect Moto would be your best bet in getting a properly thought out explanation.

It's cheap, it's fast, it gets less and less power hungry with every new version. Yet, it's still heavily criticized by some amigans, as far as I see, mostly by no other reason other for not being PPC or 68K(that is, not powering any official amigaOS hardware).

Thing is though X86 was available for the Amiga in the form of Amithlon or if you really want to press the point, on a bridge board that could be plugged into a zorro slot equipped Amiga.

[the_leander]

I'm impressed by the graphics... nearly Geforce2 speed. Run a resource lean OS like Amiga on it and you have a late 90s era gaming rig.

Only with few of the benefits.

If we can concentrate on the essentials: web browser and plenty of driver support. Photo and video editing.

Modern video editing has come a long, long way since the days of Newtec's Toaster. Everything these days is digital and in that arena I suspect AOS, even with all the improvements of MOS or AROS or OS4 would be put under severe pressure to cope with the sorts of demands that kind of software would make in today's environment.

Photo editing is probably more realistic, but then you're competing in an arena where Linux has some pretty mature and feature rich software to choose from.

You've then got the basic appliance which passes as a modern computer. Minus the resource hogging Windows.

As well as a significant chunk of it's capabilities due to it's being hampered by 80's era APIs and architecture. Appliances are supposed by nature of their requirements to be stable and unobtrusive in their function. Rigged up to the eyeballs with patches and kludges, AOS was never that.

 btw my Ubuntu 9.04 install uses about half the resources of even Windows XP.

But that Ubuntu install has out of the box more capabilities than you could get with AOS even with every patch and hack from Aminet installed and running smoothly (and good luck with that). It also has the supreme benefits of both being free (in all senses of the word) as well as well supported by vendors and coders.

There are probably more coders working just on the Linux kernel today then there were developers for Amiga software even at it's height to give you an idea of the scale of the mountain you're looking to go up.

This desire to try to move it back out into the mainstream, even under the limited guise of "appliance" type machines is simply that doesn't make any commercial sense (actually I think it makes even less sense than to try marketing it as a full blown desktop computer). There are far better suited, already mature and cheaper options around.

Amiga coming back was a great dream in 1997, which was realistically the last chance it ever had. In 2011 that same dream leaves a truly bitter taste in one's mouth. It's a retro hobby system and I'm sorry to have to repeat this, but that is all it will ever be at this late stage of the game. Minimig (and perhaps one day Natami if/when it's released) really do show the way to go in terms of the future of the Amiga. UAE for anyone else.

[the_leander]

But I run very little legacy code and I'm using my OS of choice because it provides a familiar development platform.

Fine.

Since your not responsible for paying my OS licensing fees, I'll support what developments I care about.

Wholly unnecessary.

[the_leander]

I'm inclined to think Video editing will be the easier of those two. Modern chips include hardware video encode so unless you're doing some heavy video processing you don't need that much power. The previously mentioned Pandaboard has 1080p hardware encode.

You know I thought about that initially but then considered the following: Many of those hardware encoders tend to be propriatary in nature, whilst I'm sure there are possibly some open source drivers that can make full use of them I really have big doubts that that level of support would be available to small projects such as AmigaOS.

OTOH The big problem in photo processing is you need ever more powerful hardware to handle ever more megapixels.  I bought Lightroom 3 a while back, I promptly upgraded to a Core i7 machine to run it!  Heavy processing on 21MP images requires a LOT of power and no Amiga flavour has that.

I'll happily defer to you on that.

BTW 21MP may sound like a lot but you can get 12MP in phones now.

I know about those phones, when I first read about them it left me feeling incredibly old.

Come back?  It was never big as a desktop platform.

commodorejohn's response to this was a work of art. However I'll clarify, whilst it was never a major player it did have a fairly viable userbase. In 1997 if there was ever going to be anything to change the course from micro market, hideously (to the point of being non-viable) expensive hardware to something a little more reasonably priced for tinkerers and retro fans alike, it was back in 1997.

@runequestor loved your response