Amiga A1200 graphics card discussion

[eXeler0]

Ok, so there are some (few) solutions, but they are rare and expensive.. The Blizzard Vision is AFAIK the only real graphics card upgrade ever released for non-tower Amiga 1200s.
The Graffiti doesn't quite qualify... (Even if it offered chunky display support)

We all know that even the cheapestr GPUs today offer performance that is by far beyond what the A1200 could handle anyway. Yet, there has not been any new products that would offer anything near the BVision.

For PC laptops there was at least one PCMCIA graphics card, but I reckon that the Amiga implementation of the PCMCIA standard simply doesn't work with such cards.

So my question is to the techies out there:
What would be the most sane approach to creating a new graphics card for the A1200:
*Could there be a new "ACA 1230" -style card with a GPU on the same card?
*Could the IndiVision concept evolve into something more similar to an actual graphics card?
*FPGA acellerators featuring both soft 68k CPU and some kind of SuperAGA implementation on the same FPGA?
*Any other wild ideas?

IMO, the most likely solution is an FPGA containing all the good stuff in the same chip.
Agree? Disagree? Opinion?

/eX

[NovaCoder]

I don't really like the idea personally.

There's already been a lot of talk about plans to produce an FPGA computer that plugs into the A1200's expansion port which will have on-board graphics and HDMI output. In my opinion you then up up with a new computer attached to a retro keyboard which is housed in a retro case.

I do like the idea of enhancing the original design though and eliminating some of the stupid design bottlenecks to create an A1200 as it 'should' have been in 1992 like a flicker fixer and chunky mode support (Graffiti).

I reckon my ideal A1200 set up would be:

Original motherboard, Indy Mrk2 (with Graffiti support enabled), FPGA based (020-060 compatible) accelerator (200+ MIPS) with an on-board SATA controller and 512 MB of FASTRAM. With this kind of set-up you'd be 100% compatible with existing Classic software (WHDLOAD), WorkBench would absolutely fly and you'd even be able to run the most demanding 68k software at a decent speed like Quake 2 and NetSurf (web browser).

[fitzsteve]

I agree with you nova, but a mobile radeon for rtg would be a nice addition

[IanP]

You still have a chipram/chipset bottleneck unless you move the graphics to a fast bus on the accelerator.

No graphics will ever fly if it's crippled by the old Amiga architecture.

[NovaCoder]

Originally Posted by IanP

No graphics will ever fly if it's crippled by the old Amiga architecture.

That depends on your definition of 'fly', if you mean modern speeds (60 FPS) or modern HD resolutions (or both at the same time!) then of course the original architecture will not deliver.

But how fast do we really want our retro hardware to be?

Personally I'd be happy if I could run Quake 2 with all the effects enabled at a consistent 15 FPS in 320x240 resolution on my 25 year old hardware. I can confidently predict this would be possible using a fast FPGA (running the Apollo core) even with the existing planar (crippled) AGA chipset.

If we also got a chunky mode with Graffiti then it would be even faster

[ Show youtube player ]

[eXeler0]

I hear ya, NovaCoder. I dont want an i7 + GF9800 attached to a 1200 keyboard either. (With winuae i kind of have that already)
Rather, Id like to see something that the AAA chipset (or maybe the "Hombre") would have brought to the Amiga world.
Im pretty sure that playing Quake2 @640×480 would be possible.
That's why I'd like to se SuperAGA in an FPGa.. not because it would be anywhere near modern GPUs but it would feel like an Amiga we would have gotten maybe in 1995 if things had turned out a bit differently.

[squidbass]

If it was something like the Chameleon for C64 then I think it would be a winner.

squidbass

[ajk]

Something for the A1200 which would combine the features of the Indivision, Graffiti and BlizzardVision would be ideal, in my mind.

Everything coming out of a single VGA/DVI/HDMI port of course. This would allow all existing software to be run (including the few RTG and 3D accelerated games that exist for the Classics) without the hassle of having a costly, hot, and hard to maintain BlizzardPPC+BVision combo.

I don't see any need to go significantly beyond the Permedia 2 performance level. Just something that combines all the good things, is easy to install and run, and robust.

[britelite]

Originally Posted by NovaCoder

Personally I'd be happy if I could run Quake 2 with all the effects enabled at a consistent 15 FPS in 320x240 resolution on my 25 year old hardware. I can confidently predict this would be possible using a fast FPGA (running the Apollo core) even with the existing planar (crippled) AGA chipset.

But that's not running it on 25 year old hardware, now is it?

[daxb]

Originally Posted by britelite

But that's not running it on 25 year old hardware, now is it?

And in the year 26 the old hardware dies, and then?! IMHO it doesn`t make much sense in the future to attach new hardware to the old hardware. It will become useless when the motherboard dies. Its like update your very old car without the possibility to replace the engine or wheels. 10-15 years ago a CPU+GPU+IDE(or similar) card were be cool.

When talking about FPGA I also don`t see much sense discuss about speed (cpu, video) because it is variable. Next FPGA become faster and core can be different.

As far as I know (might be wrong) Super AGA is "just" an AGA enhancement. So if the FPGA core has AGA + RTG (p96?) is there really a need for Super AGA?

[NovaCoder]

Originally Posted by britelite

But that's not running it on 25 year old hardware, now is it?

Well kinda

At least the motherboard and chipset is 25 years old

[eXeler0]

Well there are purists out there who think old stuff shouldn't be improved, but for me, its equally fascinating to see how this old hardware *can* be improved while keeping its original soul. That's why all my Amigas are 68k and not of the new PPC kind.
So although it may not make sense completely, its important for me that whatever im running is based on the original platform.

For me, the chameleon for the C64 is maybe a step too far as it can function as a standalone device and run a bunch of other soft cores..
At that point I fell I might as well run WinUAE.. but something gets lost in the transition to pure emulation...
Anyhoo.. I will welcome future SuperAGA FPGA projects, because it will (hopefully) feel as if we finally get what was coming before Commodore went belly up....

[eXeler0]

Originally Posted by daxb

And in the year 26 the old hardware dies, and then?! IMHO it doesn`t make much sense in the future to attach new hardware to the old hardware. It will become useless when the motherboard dies. Its like update your very old car without the possibility to replace the engine or wheels. 10-15 years ago a CPU+GPU+IDE(or similar) card were be cool.

When talking about FPGA I also don`t see much sense discuss about speed (cpu, video) because it is variable. Next FPGA become faster and core can be different.

As far as I know (might be wrong) Super AGA is "just" an AGA enhancement. So if the FPGA core has AGA + RTG (p96?) is there really a need for Super AGA?

daxb, sound like your needs could easily be fulfilled with WinUAE on a fast PC? ;-)

Anyhoo, the upcoming (for a while now) FPGA arcade is an upgradable stand alone product. Not exactly sure how much headroom that particular FPGA has in regard to more dranatic future core updates, but its probably not a big problem atm.

Also, the old Natami project was probably just the thing a lot of us would have wanted, but even though the new Apollo core is coming along nicely, its probably gonna be a while before we have all the parts of a Super Amiga all in one FPGA working with full backward compatibility.

The Vampire 600 v2 is just around the corner and Ive heard about plans for the other Amigas. It would be nice to see if the 1200 version could be upgraded with better graphics while we're at it. Either in the same FPGA or a separate daughter card that would be done in the future.
The expectations on the graphixs specs are incredibly low by todays standards so hopefully it wont be a problem. ;-)

[gulliver]

A simple solution for an A1200 gfx card is just to use/reuse the available Picasso design and adapt it to the A1200. This has the advantage that is a proven design with working drivers which have been already done.

The hard part is adapting the Zorro II design to the A1200 card expansion. I believe Kipper2K has bought these conectors, so it is a matter of building a passthru and buffering the required Zorro II signals at the cards interface.

Here is the link to the Picasso remake (sorry in German...):
http://www.a1k.org/forum/showthread.php?t=40955

You can take a look at this page in english with pictures taken by Matthias Heinrichs aka Matze and originally posted into his own thread at a1k.org:
http://retrocu.blogspot.fr/2014/09/s...d-project.html

A new card design will take ages to be implemented, and that is not counting driver development. So why reinvent the wheel?

[eXeler0]

Originally Posted by gulliver

A simple solution for an A1200 gfx card is just to use/reuse the available Picasso design and adapt it to the A1200. This has the advantage that is a proven design with working drivers which have been already done.

The hard part is adapting the Zorro II design to the A1200 card expansion. I believe Kipper2K has bought these conectors, so it is a matter of building a passthru and buffering the required Zorro II signals at the cards interface.

---- 8 >< ----

A new card design will take ages to be implemented, and that is not counting driver development. So why reinvent the wheel?

Ok, looks interesting, but if it was easy I'm guessing someone would be selling these by now.
Kipper2k is around here somewhere. :-) Whats your view on this? What are the pros and cons?

Also, to physically fit a board that can be used with the various accelerator designs might be tricky. And if it would only work with a very specific hardware combo it would just be another BVision..

[gulliver]

It is easy and has been done. Just a passthru for the A1200 expansion connector, so that you can still use your own A1200 accelerator, and the zorro II interface just coming out of this adapter (much like a single slot micronik zorro busboard but with no other extra features to save on cost and simplicity).

[eXeler0]

Originally Posted by gulliver

It is easy and has been done. Just a passthru for the A1200 expansion connector, so that you can still use your own A1200 accelerator, and the zorro II interface just coming out of this adapter (much like a single slot micronik zorro busboard but with no other extra features to save on cost and simplicity).

Kipper2k, could you build this thing?

[Locutus]

one problem is that it would push back the actual accelerator back a centimeter or 2, i'm not sure a blizzard060 or such would still fit after that.

[gulliver]

Originally Posted by Locutus

one problem is that it would push back the actual accelerator back a centimeter or 2, i'm not sure a blizzard060 or such would still fit after that.

You could either:

1. Create the adapter in a way that the accelerator can be placed on top of the Amiga motherboard (a U shaped adapter), and the gfx card can be placed in the area of the trapdoor (as it is smaller than an accelerator). See attached picture. Of course some rubber feets will be required to support the accelerator so it does not tilt or bend the adapter.

2. Rehouse your Amiga 1200 in a tower

[eXeler0]

Originally Posted by Locutus

one problem is that it would push back the actual accelerator back a centimeter or 2, i'm not sure a blizzard060 or such would still fit after that.

The Apollo 1240/60 is also a tight fit as it is.. I think there's a bunch of accelerators that don't have much space left. I suppose 1cm is doable, if you take the whole thing apart.

/eX