So you were put in charge of making the 060 based amiga

[fishy_fiz]

I guess all you can really do is fix the bottlenecks. A minor chipset revision to be able to access more chipram and fixes/tweaks for exisiting problems. Not to be negative, but by this time amiga was too out of date to bother fixing in its then current guise, even '060's by this point in time shouldve been bare minimum. Instead of tinkering around that time it'd have been time to start things again as amiga had too many hardware problems/shortcomings to fix. Doesnt mean I didnt/dont love the classics, its just that by the time aga was out they where so far behind that trying to fix them was a lost cause and they shouldve instead made new machines rather than tweaking technology from the early 80's.

p.s. the above is written assuming this thread is talking about a successor to a1200/a400/etc. and is based on realistic ideas. If we're going a little more into the realms of the hypothetical then some sort of 3d acceleration seems a good idea (even '060 + reasonable (even for the time) 3d acceleration could give amiga gaming a kick in the pants), 16 bit audio, useful sprite hardware (even "fixed" aga level sprites would be nice (ie. 8 useable 64pixel wide sprites)), at least 16 bit screenmodes (2x8bit playfields would be nice), the ability to mix and match playfields depending on bit depth (ie. 2x4 bit playfields + 1 8 bit for triple playfields, etc.). It's a shame Amiga hardware never went that one extra generation. With some small alterations to end up with the type of ideas I suggested (which are still very dated) my creative retro impulses wouldnt need to be so compromised.

[Buzzfuzz]

- AAA Chipset, other chips designed into 1 large chip, socketed
- Make it able to address 128MB Chip ram, fit it with 32MB and future expansion for up to 1GB in the A5000 and fit 8MB in the A1600 also have expansion port for more
- Make it able to address 2GB Fast ram, fit 128MB in the 5000 and 32MB in the A1600
- 060 socketed and hybrid bridge to x86 processor, can also be used to run Amiga OS
- Both models IDE controllers, integrated sound, network
- PCI bridge to add PCI video card in the A1600, PCI and ISA slots for the A5000
- 1.76MB 3,5" floppy drive
- Integrated CD-ROM drive in the A1600, normal 5,25" unit in the A5000
- Only 2 Zorro III slots in the A5000

[the_leander]

Draco was great for its purpose (video), but it's not a full-blown Amiga.

It may have lacked a commodore chipset, but the fact is, AAA was so far from being finished that by the time it would have been ready it would have been hopelessly out of date. Draco, using off the shelf components for graphics and sound really was the only realistic way of doing things if Amiga was to stay in any way relevant past 1995.

It was more of an Amiga than anything I've seen since C='s fall with the exception of the Minimig.

[tasmanian guy]

A 68030, 50mhz processor, with 2 meg chip and 8 megabytes ram (2 x 4 megabytes sim, leaving 2x72pin sim sockets free)
 
Workbench 3.1 would fly along with RTG graphics and TV output and VGA output as standard supporting bitplane and chunky modes.  16 bit stereo sound as standard.  Housed in a standard AT case, using standard ps2 mouse / keyboard with a 100 megabyte hard drive.  15 pin pc joystick adaptors x 2.  Price $1,000 without monitor, including Deluxe Paint VI, Word for Workbench, and Amiga Lotus 1-2-3 SE.
 
Games that this box would run without any issues:
 
Doom
X-Wing
Wolfenstein 3D
Duke Nukem 3D
Half Life
Flight Simulator Amiga 95
 
The rest they say is history!

[Hattig]

For the consumer version I would have gone with a 68030 (probably a 68EC030 for cost reasons). However I would have aimed to release an A1400 in late 1994 (for the holiday season), probably with a 28MHz 68030, 4MB RAM and a 1.76MB floppy drive, but otherwise keeping the same components as the A1200 to keep development costs down. Ship it with Sim CIty 2000 and a Doom clone.

In 2006 (for the holiday season), the A1600 would have had a 40MHz 68030 (a sad side effect of the high prices the 68060 commanded, and the high power requirements of the 68040). Remembering the cost of RAM back then, I don't know if 8MB would have been viable for the base configuration, but 6MB (4 chip, 2 fast) could have been viable. The chipset could have been updated for this revision, to include a 256 colour native chunky mode, and a full 32-bit blitter. And maybe 16-bit audio. And a VGA connector as the chipset could support higher resolutions. As a simple machine, the motherboard would have been reduced in size despite the updates, which would have reduced prices. Use of a SIMM slot for the fast RAM would have allowed an 8MB shipping product at a higher price.

[Amiga_Nut]

Well forgetting the top end model, which they could kludge together with all sorts of nice Zorro II/III cards, the base model would have to do what your average PC did.

I think my first PC was 1992, and that was 999 bucks for a 486 25mhz clone. I think by 1996 most people were on machines capable of VideoCD software playback thanks to MMX.

So you would need....

A CPU as capable as the minimum average spec of the time PC people were getting after giving up on their beloved Commodore to ever come back. Not just for games, but for doing raw filled 3D and texture mapping cheats like in Doom and Quake. So that would be a minimum of 060 50mhz (which is about Pentium 100-120mhz performance I believe as the 060 is a clock doubled CPU IIRC)

Graphics, whatever you do if you don't have a native byte per pixel mode you might as well carry on going bankrupt, super fast VRAM chunky mode for 320x256 AND 640x512. A 32bit blitter as powerful as those in the VRAM PCI cards in PCs. Rest doesn't mean a whole lot. And AAA was MORE EXPENSIVE than buying things off the shelf from people like Diamond.

Sound, 8 channel 16bit sound with variable panning, NOT 4 channels wired left/right.

A space to slot in an optional cost price CD-ROM unit like laptops of the time had.

For serious stuff, fix the bloody TCP/IP, finance a decent browser....ie slap Netscape around until they write one for you for a fee, and make sure your serial port can handle 56k modems

That's about it really, without coaxing someone like RJ Mical to do you something similar to the 3DO chipset forget anything else really, and accept that 3DO, PSX and Saturn WILL piss all over you and you have lost the advantage you had in the 80s with A500 vs Sega Megadrive/Genesis. Still an 060 and fast chunky native graphics mode would be enough to do an identical version of shitty game of the decade Resident Evil and actually a pretty nice version of Tomb Raider.

End of line.

[save2600]

Lots of great ideas already. Especially increasing the Chip Ram and having all major chips be socketed. Oh and they should have done away with this 15khz nonsense after the A3000. Built in accelerated graphics too ala CyberGrafx/Picasso or maybe a compatible AAA design. 16-bit sound as standard and call her Super Paula.      

I'd also make the case larger and a bit less cramped. Should have more personality to it than the A4000 box. Something alone the lines of the A3000, but maybe with standardized 3-1/2" HD floppy drives and a slim line CD-ROM, like what the laptops had.

Speaking of laptops, I would have done away with the form factor of the A1200 altogether and gone with something more along the lines of the A600. This machine could be 100% surface mounted, unlike its big box cousin. And I'd design it so that the customer would have the option to turn his new Amiga into a portable by adding a battery and a screen that you'd simply thumb-screw into place on the back of the machine. If the battery technology still sucked and you couldn't add something internally, then a real thin, but large footprint designed battery that mounted underneath (raising the machine a bit too for typing) would have been implemented.

So yeah, I'd have a mid to hi-end "workstation" type big box and a semi-portable, which was what the A500/A600/A1200 kind of where or could have been. If someone wanted a nicer keyboard for their new portable - there'd be a port so you could hook a regular/large big box styled keyboard to it.

Both machines would feature the 060 (cost would be driven down thanks to aggressive negotiations), a true accelerated graphics chipset, 16-bit sound and HD floppies. Finally, I'd market the hell out of the machines in professional circles as what they were 'and could be' instead of narrowly focusing on a particular market like they did so often with graphic artists or other "creative" types.

Cut deals with schools on the hardware and make up for it in licensing and software.  'Power Up' programs for consumers would be mailed out to each and every Commodore owner and speaking of lists/serial numbers - I'd have maintained a much better relationship with previous Commie owners all this time. Newsletters with constructive input forms for sure instead of paying some company to tell us who are customers are.  lol

[the_leander]

So that would be a minimum of 060 50mhz (which is about Pentium 100-120mhz performance I believe as the 060 is a clock doubled CPU IIRC)

Just a small point since the rest of your post isn't that bad, at 66Mhz the 68060 would kick out ~88Mips peak. An original Pentium clocked at 60Mhz did 100Mips.

Also consider that the maths performance was better on that original pentium even before the introduction of MMX.

[Iggy]

Offer either an '030 or '060 via an expansion slot that can also accept PPC or PPC/060 combo processor card.
Forgo complete graphics compatibility and go RTG ( unles RTG slot can be made to coexist with an older graphics system).
Limit Zorro slots and use PCI slots for primary expansion.
Consider another X86 cooprocessor card that is better integrated into AOS.

Offer a tower and a unified desktop model (like the 1200). The later may not be able to have all the features listed.

Drop AOS in fovar of MorphOS.

[Zac67]

Big box:

Dump Zorro, PCI only. 5 slots minimum. If it's very low additional cost include 1 or 2 Zorro slots.

SCSI was somewhat expensive, make it optional. EIDE (MW-DMA!) as standard.

Separate CPU & RAM card, no RAM on mainboard. Memory map allows up to 1 GB of RAM on the CPU board.

Move graphics chipset to PCI/Zorro card to make it optional later on.

[Amiga_Nut]

Just a small point since the rest of your post isn't that bad, at 66Mhz the 68060 would kick out ~88Mips peak. An original Pentium clocked at 60Mhz did 100Mips.

Also consider that the maths performance was better on that original pentium even before the introduction of MMX.

I think I remember the 060 being clock doubled internally, but couldn't remember the exact performance between 060 <> Pentium. Still 060 is best we had before going PPC so not much choice really.

My point was gaming was going 3D with PC and PSX/3DO/Saturn and we needed something similarly powerful even in a base model and an 030 or 040 wouldn't cut it as custom 3D chipset was impossible from Commodore in 96. Even if base bargain PC in 96 was P75. Either that or licence PowerVR or something like that but I think 96 is too early for that kinda hardware really.

edit: Possibly a DSP too, not too sure how well the one from the Atari Falcon would work with an 060 doing 3D games acceleration, maybe good for matrix manipulation for polygon game engines.

[Amiga_Nut]

Big box:

Dump Zorro, PCI only. 5 slots minimum. If it's very low additional cost include 1 or 2 Zorro slots.

SCSI was somewhat expensive, make it optional. EIDE (MW-DMA!) as standard.

Separate CPU & RAM card, no RAM on mainboard. Memory map allows up to 1 GB of RAM on the CPU board.

Move graphics chipset to PCI/Zorro card to make it optional later on.

Having Zorro slots would mean you can cut a deal with Sunrize for 16bit sound saving development costs, ditto for 24bit graphics card with much higher resolutions for photo manipulation like in Photoshop or high resolutions for DTP and 20" monitors etc etc. Add a PPC and Blizzard3D and you are rocking for peanuts dev costs. This would mean very little updates to WB needed as the drivers are there already rather than the PCI cards.

[Amiga_Nut]

What you could also have done was to fit 28mhz 020 and 1mb fast 32bit RAM in the A1200 to replace the A600 type entry level system in 1995/6 too. This would be good enough for basic Doom/Wing Commander type games if programmed properly.

Again make a standard slot for A1200+ and 'A1600' to take a CD-ROM module sold at cost price to encourage competition with PC-CD market too.

[the_leander]

My point was gaming was going 3D with PC and PSX/3DO/Saturn and we needed something similarly powerful even in a base model and an 030 or 040 wouldn't cut it as custom 3D chipset was impossible from Commodore in 96. Even if base bargain PC in 96 was P75. Either that or licence PowerVR or something like that but I think 96 is too early for that kinda hardware really.

Your choices for 95/96 were S3 Virge, Matrox Mystique, ATI Rage or later in 1996 the Rendition Vérité V1000 or the 3dFX Voodoo pci.

PowerVR were around in 96 and they supplied Matrox with their m3d parts.

However for an all in one solution your best bet was probably the V1000.

edit: Possibly a DSP too, not too sure how well the one from the Atari Falcon would work with an 060 doing 3D games acceleration, maybe good for matrix manipulation for polygon game engines.

Why not simply have it deal with the sound?

[minator]

So lets say Commodore had built an 060 based amiga, before jumping ship to other hardware (or dying).
You were put in charge of the design plans:

You couldn't dump the chip set for lower end machines because of compatibility and the top end machines needed it for video.

I'd release the AAA/DSP combo they were already working on.  Wouldn't matter if it didn't have the best graphics, but the DSP alone would have blown everything else away.

That would keep things going until the next chip set could be developed, with 3D of course.  You could put a chipset like that on a card and sell it as a PC graphics card - like SGI should have done but never did.

As for the OS it would have to be ported to something better, I would have looked at Alpha but it probably wouldn't have been financially possible (they were very big, expensive chips).  Amiga would have almost certainly have gone PowerPC.

It would also be apparent the OS needed upgrading, for that I'd buy Be inc.

[Nlandas]

I'd have put Dave Haynie in charge and gone to have a cocktail.

I suspect if Dave had been given more control over the direction of Amiga hardware, he'd have started to adopt industry standard components/standards. Perhaps, with backwards compatibility in a few models.

One of my favorite comic books series was What if......
E.g. What if.... Wolverine had killed the Hulk.

So, I propose... What if.... Dave Haynie had full control.