ARM based Amiga?

[SamuraiCrow]

Aros is already supporting ARM... and AmigaOS will propably never be ported to other platforms because of licensing issues

+1

[SamuraiCrow]

I stand behind my teammates on the NatAmi project.  The N68070 will kick butt when it comes out.  The N68050 will be the best softcore until then once it comes out.

I used to use an 800 MHz Pentium III with the basic performance of a Raspberry Pi.  If you don't need it to fit in small places I'll sell it to you gutted with parts removed for $5 + shipping.  

What made Amiga fun to work with was that you could tinker under the hood with the Hardware Reference Manual as a guide.  No need for drivers because they all had the same chipset.  No need for emulation layers because they all had the same processor series.

I personally think the next stage is a compile-time VM that uses AROS as the runtimes.  Then it will run full-speed on any platform that AROS runs on.  This included hosted AROS on Linux, Mac, Windows, Android 2.2+, PPC Linux, and ARM Linux.  The catch is just designing everything to fit together.

[SamuraiCrow]

@Tripitaka: I've been thinking on similar lines for some time in fact, but now I've heard about Denver it seems a good deal more plausible.  We've got a Beagleboard in the house, I wonder if we could get it to play nicely with a natami/minimig somehow...

In order to get MiniMig to play nicely with a BeagleBoard, you'll need a field programmable gate array chip to hold the graphics and sound core.  I don't think the BeagleBoard comes with one.

The NatAmi is an FPGA-based SoC that will have both the CPU and graphics and sound cores all on one FPGA chip (someday, anyway).  About all that we have in common with the Denver chip is that we're pursuing some of the same market space.  I'd love to have an up-to-date chip fabrication process on a full-fledged FPGA but that's not the priority now.  Getting the bugs worked out is the priority.

[SamuraiCrow]

Nice, an ARM 'Amiga' Not wanted, because Amiga is hardware, not software (680x0+chipset). My Amiga hobby is based on actual Amiga hardware, and not some non-Amiga hardware. The software just makes the machine useful. What Amiga really needs is a better OS (written from scratch, and for 680x0+chipset) and new software that's also written from scratch.

I disagree about the OS sucking so bad.  It's the compilers that suck.  If we could get an up-to-date compiler for 68k, we could write software in C, C++, and (by extension) PortablE that didn't suck instead of twiddling around in Assembly for our larger apps.  Also, since AROS is written in C we'll need the compilers to work better for that as well.

[SamuraiCrow]

@Thorham

The reason Exec and the other libraries are stuck in the past is that they were written in 68k Assembly.  That just made them harder to bring into the 21st century when 3.5 and 3.9 were written.

As for the compilers, LLVM will soon be able to use a PBQP register scheduler that can store multiple small variables in one register.  This would improve register loading and make it look more like somebody sat down and hand-assembled the whole thing.  The GCC compiler could do similar things but doesn't because the 68k backend is so antiquated.  Nobody will put the time nor energy into GCC 68k because it's thought to be a dead architecture outside of embedded controller use.

[SamuraiCrow]

Haiku is the best OS for the PC.  But there's no software for it.  I agree with Nicholas on the hosted AROS port to Haiku.  That would allow lots of AROS software to run on Haiku.

As for the 680x0, it's only as much of a dead end as the NatAmi requires it to be.  We'll be bringing up-to-date features to the N68050 but it will still be a softcore for quite a while.

@Thorham
While the 68020 is capable of better OS functionality, I doubt that it's going to help.  Most of the extended addressing modes are slow on the '020 and '030.  The '040 and '060 fixed that at the expense of some hardware compatibility particularly for floating point.

Also, memory protection would slow the real Classic Amigas to a crawl.  Believe me, it's not needed.  A better solution would be to build software on top of the PNaCl (Portable Native Client) sandbox that is going into the latest Chrome web browsers.  It will allow 97% runtime speed compared to standard native compiled code on x86 and still be secure.  All this without memory protection.

[SamuraiCrow]

Two problems: 1) This doesn't run on Amigas, 2) What fun is it to write for a browser?

1)  It will run on AROS.  2) It can be hotwired.  

My definition of Amiga is hardware, namely chipset+680x0, and sadly this is dead until someone buys or licenses Amiga hardware and starts building the things, and that's most likely not going to happen

Perhaps you've never heard of the NatAmi nor the MiniMig?  There are some flaws/shortcomings in the Amiga chipset that need to be addressed and so a total backward-compatible rewrite of the chipset design would actually be preferred.  Case in point:  The DMA scheduler doesn't take advantage of sequential memory bursts so it's going to need to be changed in future hardware.  Fortunately, the NatAmi is being redesigned to use DDR2 memory instead of fast-page and does exactly that.

And yes, I want a new Amiga, but no one will make one I'll probably get an alternative machine after Amigas die physically (hopefully that won't be for another 10+ years...).

http://www.natami.net

[SamuraiCrow]

@Mrs. Beanbag

Would a conventional Kickstart with ROM switcher built-in do the trick for you?  The NatAmi has that already.

DVD playing is a problem due to region-locks on the copy-protection.  Also there are licensing issues on the copy-protection as well.