Does eight bit computing excite you?

[ChuckT]

Can a market for 6502 Computers exist anymore?

If you were asked to develop material for learning, what would you say?

If you attached an IBM video chip to a 6502, would it be a worthwhile computer to get involved in?

I was thinking that it would be easier to develop more eight bit systems into something that the Commodore 64 never was but I think a lot of people are hooked on game cube and x-box.

[trekiej]

A high speed 8 bit system with nicer graphics would be great to see.
Higher resolutions could make it more atractive.

[ChuckT]

There are newer processors which can handle twice the information at roughly (I haven't exactly looked it up) the same speed and the UARTS can be interfaced with SD cards.

I talked to someone in the industry but I haven't decided what to do because few companies can compete with the price Commodore was selling the C-64 for so I am in the dark as to what the final cost will be but I believe it will be more expensive than a cheap PC.

I think that more people would be familiar with some of the older technology, that it would be easier to learn and programming could come back to some extent.  I think the extra bits and newer processors may create more architecture for people to learn which increases the difficulty level.

[persia]

Are we talking something like an old heath kit project?  What would you do with it once it was built?

[Speelgoedmannetje]

ChuckT wrote:
I think that more people would be familiar with some of the older technology, that it would be easier to learn and programming could come back to some extent.  

Yes and no.

Considering assembly, simple tasks are much more of a hassle to program than 16 or 32 bits procs.
And more instructions means one can (almost) program in assembly as if it is a higher programming language (like C).
Which is more readable, and easier to be kept readable.

Cons are, that more complex instructions means that more actions are being done 'underwater'. You just need to learn these characteristics and then that system is 'yours'. :-)
Personally, I'd like to play/program with the Playstation 3 + Linux a bit, to see what those 7 cores are capable of :-D

[ChuckT]

You can write programs for it but Sony may be blocking some aspects of the Sony Playstation 3 (3D Access).

http://ps2dev.org/

Is it worth writing programs for those machines if hardware manufacturers are going to do that?

Is it worth going back to 8 bit?

Speelgoedmannetje wrote:

Personally, I'd like to play/program with the Playstation 3 + Linux a bit, to see what those 7 cores are capable of :-D

[Speelgoedmannetje]

ChuckT wrote:
Is it worth writing programs for those machines if hardware manufacturers are going to do that?

NO!

But anyhow I want to try out that Cell processor.
I heard there was a new one upcoming, maybe there's a way getting it to work with in a pc as cpu or accelerator in some way.
I just don't want x86 (or those 64 bits derivatives)

[bloodline]

Speelgoedmannetje wrote:

I just don't want x86 (or those 64 bits derivatives)

What exactly is wrong with the x86 and x86-64? I downloaded the manuals from intel last year and read through them... these are nice chips!

[Speelgoedmannetje]

I'm sorry if I have to be a bit vague here (it's quite a couple of years ago), but I recall there was something about this cpu design that forces you to juggle alot with registers and memory allocation (when programming in assembly).

[bloodline]

Speelgoedmannetje wrote:
I'm sorry if I have to be a bit vague here (it's quite a couple of years ago), but I recall there was something about this cpu design that forces you to juggle alot with registers and memory allocation (when programming in assembly).

Yeah, the x86 ISA is a bit weird, it's rather RISC like in a weird CISC way :crazy: The hardest detail is that the Opcodes and Addressing modes are not always orthogonal, as we are used to in the 68k... and there is the issue of only 8 registers, some instructions have certain register requirements too... which is a bit of a pain, but the SIMD, MMU and cache control is all very nice and lets face it only very rarely does one ever go to ASM now...

The x86-64 ISA has been nicely cleaned up, plus it has 16 general purpose regs like we are used to with our 68ks :-D (actually better than the 68k since there is no Address/Data separation) Probably the most difficult thing for a 68k coder on the x86-64 CPU is the lack of addressing modes... but all modern CPUs avoid complex addressing due to their getting in the way of high performance.

Though if you really want to suffer with asm in 2009... the ARM is really simple and very nice :-)

[Speelgoedmannetje]

It's not so much as suffer with asm code, but more of a difficulty for compilers. Compilers already make, so to say, sub-optimal binaries, and these flaws in CPU design really doesn't make things better. I think lots and lots of cpu cycles are being wasted.
Don't get me wrong, I really like backwards compatibility, but I rather see backwards compatibility as expansion rather than being used as a base.

[bloodline]

Speelgoedmannetje wrote:
It's not so much as suffer with asm code, but more of a difficulty for compilers.

Compilers don't care... The ISA can be bizarre and confusing, or neat and elegant... it makes no difference to the compiler at all. But as a Human I find 68k much easier than some horrific compiler oriented ISA like PPC for example.

Compilers already make, so to say, sub-optimal binaries, and these flaws in CPU design really doesn't make things better. I think lots and lots of cpu cycles are being wasted.

I doubt a human could make code as efficient as a compiler on a modern CPU... Modern CPU's have hidden register renaming schemes, instruction scheduling, caching policies... the list is endless... I doubt you could know all the optimization rules now!

Don't get me wrong, I really like backwards compatibility, but I rather see backwards compatibility as expansion rather than being used as a base.

The ISA is irrelevant with regards to performance, other than the big things like the removal of difficult to optimize instructions and addressing modes...

Read: :-)
http://arstechnica.com/cpu/2q00/x86future/isa-future-1.html

[persia]

I haven't seen assembly language in 20 years, since most computer chips are multicore nowadays does that make the assembly language more complicated?

[bloodline]

persia wrote:
I haven't seen assembly language in 20 years, since most computer chips are multicore nowadays does that make the assembly language more complicated?

The asm code of a multicore chip is not more complex... but writing applications that take advantage of multiple cores is much more complex... threading is a difficult topic at the best of times..

[Speelgoedmannetje]

bloodline wrote:

Read: :-)
http://arstechnica.com/cpu/2q00/x86future/isa-future-1.html

As far as I can read from that article, they state that with the virtualisation of the x86 since the introduction of the Pentium Pro (a feature I thought only the Transmeta Crusoe could do; the only difference is the Crusoe is a titbit more flexible), Intel dealt with the hardware limitations of x86, but this article didn't wipe away my concerns about an obsolete instruction set crippling compilers.