ARM for the future?

[nicholas]

It is the only choice for most computer users.  That does not mean it is the best.

Interesting from TFA (even if its shmoo):

Denver frees PCs, workstations and servers from the hegemony and  inefficiency of the x86 architecture.  For several years, makers of  high-end computing platforms have had no choice about instruction-set  architecture.  The only option was the x86 instruction set with  variable-length instructions, a small register set, and other features  that interfered with modern compiler optimizations, required a larger  area for instruction decoding, and substantially reduced energy  efficiency.
 Denver provides a choice.   System builders can now choose a  high-performance processor based on a RISC instruction set with modern  features such as fixed-width instructions, predication, and a large  general register file.   These features enable advanced compiler  techniques and simplify implementation, ultimately leading to higher  performance and a more energy-efficient processor.


I remember when RISC was going to be the next big thing before the Pentiums came out.  It is funny to see that it has taken another 15 some odd years to come back to this.  x86 is good because that is what has been crammed down our throats by a business model bent on dictating to us what we need.

Bring back the Alpha!

[nicholas]

Now that's funny!
RISC was supposed to be big. PPCs,, I64, Alpha, SPARC, all tried to push this forward.
Will ARM succeed where others failed? Who knows. But no one forced X86s on the market. Apple could have continued to solder on without the switch.
This situation has occurred thanks to continued development that has allowed the X86 to have a performance edge while maintaining a low cost.
The phenomenon is purely market economics in action. No great corporate conspiracies are involved here.

IIRC one of the Draco models had a slot for an Alpha CPU.

[nicholas]

For a hobby system it would make a lot, lot more sense. Imagine paying 500, or even 1000 $ instead od 2500 or 3000 $ for your next AOS4 machine.

In fact, I can honestly say even I, a known anti OS4/MOS/AROS/OS 3.1 person:), would consider buying such a machine because I could justify the expense.

If you are anti everything that is considered by the community as "Amiga", then why in September of 2010 did you join this Amiga community forum?

You are not "known" to us at all.

[nicholas]

Well, let's just say I'm a amigan but of different breed to the current amiga community over here.

So who are you then?

[nicholas]

Most of that is still mostly consistent with what you were replying to: A small number of "almost but not quite" general purpose registers for integer ops and logic.



... but only for 64 bit code. And you still carry along the whole messy old legacy. x86_64 is *much* better, it's almost tolerable. But it's still far uglier than m68k from an assembly programming point of view. Not that that matters for most people these days, though, since there aren't many of us left that actually write any reasonable amount of asm (I mainly do for my compiler backends).

Everything is ugly compared to 68k assembly.

[nicholas]

I still have a soft spot for 6502 assembly... It's beautiful for its simplicity. Then again, of course it can't compete in actual usability.

It definitely has it's own charm, but then again I have a weird pseudo-masochistic fondness for 8086 real mode assembly. :lol: