Amiga Coldfire project dead?

[Amiga_Nut]

I was looking around the net for info on a coldfire project for Amiga, and I found the following website...

http://www.cdtv.org.uk/coldfire/

However the last update on there is for 2004...it's 2010!! So what happened? Did they give up on it 6 years ago?

Coldfire V4 is an interesting CPU, at 266mhz it runs at a nice and efficient 400 MIPS.

Anyone have any news to shed light on this sad state of affairs? It's not a solution to a next generation Amiga without someone tooling up to put AGA chipset on motherboards I know but it is an interesting CPU and accelerator route, and given Coldfire CPUs cost bugger all it's a shame this went a bit quiet half a decade ago.

(I am not interested in what people think of Coldfire as a solution (or any other CPU you believe is personally better), only information regarding the project being abandoned or another group taking over this task in some small way)

[Amiga_Nut]

Thanks for the link. I totally understand though, working full-time to cover your life expenses leaves very little time/energy for doing stuff like this.

If I won the lottery I would put up a £200,000 bounty to finish this.

(£200k is including combining hardware design with a Kickstart work around compatible with the v4 Coldfire CPU)

[Amiga_Nut]

If it was combined with a Kickstart work around project it would be great and a re-compile of Workbench (Amiga Inc being contacted for a licence I guess) then there is no reason you couldn't go native with a Coldfire V4 setup...and V5 when it is produced in the future. And let's face it not much software outside games requires a 100% 68k inefficient mode anyway.

Remember I am talking about a plug-in board that doesn't cost a fortune like current PPC boards for classic Amigas on ebay.

I didn't want to get into the alternatives debate but we only have 4 CPU choices left.

ARM = total rewrite of everything from scratch.
X86 = total rewrite of everything from scratch + a million drivers for all x86 hardware combos
PPC = hugely expensive hardware so no sale to people with £50/100 machines
Coldfire = not a simple transition but maybe the simplest of the 4 in reality?

400 MIPS is a lot of performance, in the short term it's not possible to use all that performance sure. But it is CHEAP technology, and cheap = economically viable business = profits plowed back into using more features of the actual Coldfire architecture via modifications to KS/WB. Coldfire is also being actively developed unlike PPC which is the only other CPU remotely compatible with 68k systems in any way.

Maybe a future solution would be 266mhz Coldfire v4 + Minimig style FPGA for OCS/AGA chipset combined in a single desktop machine. But only after upgrade cards for 500/1000, 1500/2000, 1200 and 3000/4000 models were released. This would cater for a new bread of Amigans who never lived the time of Amiga.

All I was interested in was an alternative to £700 PPC cards on ebay that don't run classic Workbench 1.3/Lotus 2/Rocket Ranger natively anyway and sit there like expensive co-processors. Even if it runs as 'badly' as a 68060 clocked @ 66mhz who cares in the short term anyway if it costs less than a 060 card for 2000/4000/1200s?

[Amiga_Nut]

Get a scratch card if you win a tenner, I'm still gonna hold you to the above statement

Also, the 68k AROS project will go a long way to solving the kickstart issue... The Amiga software issue aside...

I will put up the same 20% as a bounty from my winnings, so that's £2 :p

[Amiga_Nut]

EmuTOS.....that's no small achievement in itself though given that the original TOS wasn't anything other than 68000 compatible, no other 68K CPU could run the ATARI ST OS at all (which is why the fastest STs ever were 16mhz 68000 systems).

I'm sure if £200,000 was on the table as a bounty then it would not be impossible to get Amiga Coldfire into reality...that's like 10 years of a very decent salary (salary left after taxes that is) or 20 years of your typical salary to be honest

Of course I would retain all the rights and lock it up in my castle I have built with my lottery millions

[Amiga_Nut]

Well, nothing on that page indicates that. Care to point out any benchmark results in running some actual 68k code?

Have a look at the ST/TT section's Firebee thread on atariage.com/forums/

Might be someone with technical info or some more leads to other sites there.

[Amiga_Nut]

Well, you could as well just ask them how they're going to handle any code that depends on:

• 68020 multiply/divide instructions behaving like they do on 68k
• MULU and MULS setting the overflow bit
• ASL and ASR setting the overflow bit
• MOVE.B ,-(A7) and MOVE.B (A7)+, changing ea by 2

As far as Atari TOS goes you only deal with pure 68000 and nothing ese. Even 68010s will crash GEM/TOS btw.

[Amiga_Nut]

PPC = total rewrite of everything from scratch + a driver for all ppc boards/chipsets
ARM = total rewrite of everything from scratch + a driver for all arm boards/chipsets

I really don't see the difference with x86...

Because there are millions of motherboard/graphics/audio/hdd controller/network card combinations even in just the last 5 years. And that doesn't include all the laptop models. The mind boggles!

PPC via stuff like Warp OS is much further. ARM has 100s of combinations at best and PPC not much more either.

This is why Windows is such bloatware probably. But Windows has 70% drivers from previous releases completed anyway most of the time.

[Amiga_Nut]

Why not build hardware-based dynamic recompilation (or translation) into the execution pipeline, i.e. an MC68000 compatible ISA that translates only the necessary bits to the ColdFire ISA (the "microcode," so to speak). Could such a solution be faster than a complete recreation of the MC68000 ISA in an FPGA? Or could a solution that translates from one ISA to another, e.g. MC68000 to ARM, be just as fast or faster?

The inefficiency issue still gives you 060 performance whilst KIS.

I was really wanting just classic OS ie WB 1.3/2/3 I don't want it for OS4/MOS/AROS. An 060 used for an A1200 costs what £300 used? And 060 speeds on A500/1000 are not available full stop.

This is all I want.....

Load OCS/AGA game and play it without chipset emulation
Load Dpaint and speed up the calculations
Give me max ram specific version of OS allows.

Coldfire is cheap tech and has some semblence of 68k compatibility. As stage 2 by all means recompile KS/WB 1.3/2/3 to run in native Coldfire v4 at 400 MIPS.

As for £200,000 reward I offered hypothetically not being enough, I am not hiring anyone or paying you to do a project. I am stating merely if anyone can provide me with a working Coldfire accelerator and suitable OS patches for an A500/1000/2000/4000/1200 I will purchase them ALL for a one off fee of 200k. As a condition of the sale I wish to have the IP transferred to me for free.

So all I have bought is 5 Amiga accelerator cards for the princely sum of 40k each and the taxman can sit on his finger and swivel, nothing to do with him just like if I bought 100000 modified 75mhz 060 accelerator cards off you second hand. The IP transfer is to ensure my purchase is unique as far as the taxman is concerned. If I then choose to employ you in a business venture using my newly acquired IP and patents etc only after this point is there any income tax surely, not from the sale of the fruits of some hobby activity I found interesting enough to purchase exclusively for a large sum of money. If you had a McLaren F1 super car I would purchase that too from you lol, this is no different.

Besides half of the UK IT professionals are on the dole in the UK thanks to two successive idiot governments running the country into the ground without remorse

[Amiga_Nut]

Carl Sassenrath discusses this in one of the Amiwest videos IIRC.

Unbelievable!

I am surprised the Falcon runs any ST programs at all, games or serious software. But as far as I recall ST GEM and TOS where pretty much rushed and GEM was a hack from x86 DR DOS version source to 68k in record time.

[Amiga_Nut]

This train of thought always baffles me.
Those same nic/gfx/sound/scsi/sata/etc. cards are also options for other architectures. The only thing that's specific to an architecture (besides the arch itself) is core logic (and even that isnt arch dependant all the time). AROS has done fine over the years on multiple north/south bridge combos without needing specific support for them (as is the case with modern computing), why would it be different for any other OS ? If x86 isnt an option due to the sheer volume of options then either is arm, ppc, 68k, etc systems with agp/pci/usb/etc... too many options  

PPC laptops don't exist, maybe there are 3 different ARM mobile motherboards and a handful of hobbyist ARM motherboards with no sound/video/network card on PCI/AGP.

How many actual drivers are there for PPC boards with PCI or AGP? Nobody will be buying anything like all of the 1000s of different video cards that may physically slot into a PPC board if there is no OS4 driver really.

Think of the Mac, the largest group of PPC users in the past, yeah sure they had AGP or PCI slots in their machine but very few cards had OS9 or OS X drivers for them in the box so Mac users didn't buy them and so the subset is tiny, and sound/network/motherboard is locked per generation of CPU pretty much. Only PCs can have 100s of motherboards with 1000s of graphics cards 100s of network cards and 100s more sound cards plugged into them and then you have totally bespoke laptop hardware...a particular Toshiba XP CD has drivers that the Dell XP CD does not. And the retail copy of XP would have drivers for neither. I have lived this nightmare for a decade and I know it only too well trust me

[Amiga_Nut]

I'm wondering how does the Falcon work then

I'm not sure how many ST games (ie not GEM based application software) work on the Falcon, but TOS on the ST is just CP/M written specifically for 68000 CPU which is probably why it won't work with 020and up. And CP/M is from Commodore PET times and the port is very rough and ready for 68k.

ST+TT+Falcon all have GEM on top, but I think it is completely different OS underneath on Falcon and TT ie not CP/M 68k.

TT runs bugger all ST stuff except GEM based apps if I remember correctly and you couldn't burn a ROM using the Falcon/TT equivalent of TOS and place it in an ST as it wouldn't work (and they probably aren't small enough to fit in the ST's memory map probably...~160kb IIRC).

[Amiga_Nut]

Some Apple Notebooks running MoS

If I meant Mac I would have said Mac, I meant PPC as in OS4 compatible or any embedded controller type setup people might want to make useable with OS4/AROS/MoS ie non-Apple Mac hardware.

But even if you include Apple notebooks, compared to the entire range of x86 Windows targeted notebooks during the lifespan of 680x0 and PPC based Mac notebooks it is still a tiny drop in the ocean in combination of hardware drivers required to work so a rewrite of KS/Wb 1.3/2/3 is not feasible AND anyway MoS on Macbook won't play Rocket Ranger natively as there are no AGA/OCS chipsets in PPC Macs so it fails the requirement

Like I said I am only talking about a Coldfire accelerator card solution and modified KS/Wb on classic ie real Amigas from Commodore not all this OS4/MoS stuff on Apple or other PPC hardware with no actual Agnus/Paula/Denise in sight on the m/b and no way to load up original floppies.

If I am wrong then I suspect you are 90% complete with an x86 version of MoS and native drivers for every notebook and PC ever sold from 1999 to 2010? lol

[Amiga_Nut]

@Bloodline

Indeed, DSPs helped that crappy 3mhz Super Nintendo CPU to do games like Star Fox/Pilot Wings or even the 68000 in Sega's consoles do Virtua Racing.

Never mind A1200 the bloody A4000 should have had a DSP as standard given the price of the machines IMO.

[Amiga_Nut]

You're right, there's no such hardware if you exclude Apple. But I wonder why you would do that?

Apple PowerPC hardware is still the best there is, and it has great support and repair services. Why go for some rare (tiny production runs), prohibitively expensive custom HW with substandard support when there are better options around?

Try to get support, spare parts and/or repairs for AmigaONE SE/XE or Micro-A1 today and you get the idea.

I excluded them because classic original workbench isn't being ported to Macbooks or iMacs, ie Workbench 1.3/2.1/3.0 and you can't plug them into an A500 as an accelerator card.

I started the thread to try and find out if/when Amiga Coldfire project died and we all got sidetracked on to the age old "I want OS4 for X86"  or "I want PPC OS for Amiga" debate but all I was pointing out was that x86 is no standard at all due to millions of component combinations/options so if you are porting an OS AND writing the drivers for everything defined as either PPC or x86 then PPC is a smaller mountain to climb really even with Mac PPC hardware. Anyway ARM or x86 accelerator for classic Amiga is even more impossible than Coldfire at the moment

But as I stated I am only really interested in this potentially cheap tech to accelerate actual original Amigas not replace OS4/MOS motherboard projects like SAM. The dream is you plug in a card that costs less than an 060 and gives you at worst 060 performance inside real Amiga  like 500/1000/2000 and 1200/4000. You can't even get 060s for A500/1000 and 2000 cards must be incredibly rare.