New type of accelerator design?

[Damion]

and pretty sure there was a 80 mhz 040 for Apples.

This was marketing, as by design the '040 internally ran 2x bus frequency. A lot of amiga '040 cards advertised at 40MHz also ran 80MHz internally (TekMagic, Cyberstorm, Apollo, etc), but 80MHz bus simply wasn't possible.

[Karlos]

A 68020 compatible core (possibly with a few 040 user mode extensions such as move16) implemented in an FPGA would seem the best bet to me. I don't know how feasible that is, but if it could be done would be an ideal accelerator base maybe. If it could be implemented, along with a basic memory controller that allows generic DDR memory (simply based on availability) that'd be nice.

Something like that would certainly have me interested

[amigadave]

A 68020 compatible core (possibly with a few 040 user mode extensions such as move16) implemented in an FPGA would seem the best bet to me. I don't know how feasible that is, but if it could be done would be an ideal accelerator base maybe. If it could be implemented, along with a basic memory controller that allows generic DDR memory (simply based on availability) that'd be nice.

Something like that would certainly have me interested

Me too!  I think it can be done, and will very likely be completed by someone, or some group in the near future (1 to 3 years).  The longer it takes to be invented, the less people will still be around here and interested in paying for it, as the Amiga community continues to splinter and disintegrate.

[Karlos]

Regarding the original suggestion, it's hard to see how using a "fast and cheap" x86 makes sense when you can just run UAE to the same end and have less latency from the rest of the system.

[lapeno]

SDRAM and later memory modules became 64 bits wide which complicates a bit the bus design between the 32 bit processor and the RAM. It is interesting however that the operation of synchronous RAM (e.g. SDRAM) is much more simpler than an asynchronous SIMM's. SDRAM would be an ideal choice for the accelerator but this 64bit wide operation must be solved somehow. I don't want to waste half of the capacity of one module just because the design is simpler that way.

[rkauer]

Why not? SDRAM modules are common as mud and the 32Mb (using a 64MB module) is enough for Amigas!

[blanning]

Some time in 2010, Jens will release the Chameleon for the c64.  If you're of the attitude that the upgrade makes the main machine pointless, then the chameleon is ridiculous.  It's basically a computer by itself and, iirc, can even run as standalone c64 emulator without even having a c64 attached.  It has ps/2 and vga ports.  And it can run the c64 in various accelerated modes, or not at all.

I'd like to see something like this for the 500/2000 and the 1200.  You plug it into either the 68000 socket, cpu socket, or the 1200 expansion port (ideally the same board doing all three with adapters if that's even possible).  The board would have an FPGA (I think, i'm not a hardware person) that emulates any 68k chip or even a powerpc.  Ideally, it could emulate the AGA chipsets for the 500 and 2000.  It would include ide and scsi ports.  A CF hard drive could be supported through those adapters.  It would include two clock ports and maybe ethernet.  USB could be supported through the clock port with a subway, or more ideally, just built in for 2.0 support.

Of course, 128mb or 256mb of ram.

If you really wanted to get crazy, instead of including ide or usb on the board, just include a zorro 3 slot.  I bet you could squeeze a deneb into a 1200 that way, or a buddha board for a 500, parallel to the motherboard.

RTG, vga, ps/2 mouse and keyboard, and catweasel-style support for 1.44mb floppy drives could even be included.  Although some of this could be handled through indivision and lyra type upgrades.

The board could support all kinds of options through the boot menu.  Choose a CPU, chipset, and whatever other options.

In the end, it would be a ridiculous upgrade, just like the chameleon.  But the user could choose any configuration they want, yes, just like WinUAE.  But WinUAE still wouldn't be an amiga.  To me, the ultimate hardware upgrade would be more fun than WinUAE.
   
brian

[Tron2k2]

That would be HOT. I'd love to plug such a thing into my A500.  I always wondered why someone didn't just make a board which went into the 1200's accelerator slot, had a USB2 cable at the other end, and essentially made the 1200 a slave to the modern machine on the other end.  USB2's bandwidth exceeds anything happening inside the 1200, it could just see the other box as a really fast accelerator.  But at that point, why not just do a little Atom board on a trapdoor slot size form factor?

Even looking at the Efika's 700mips speed-that's more than twice a Cyberstorm's oomph, plus it has PCI.  Someone could make a CPU board based on its design, with the hangup being big box Amiga CPU slot>PCI interface.  Then you'd even have all those little Efika extras-legit ethernet, IDE, decent sound-the gang's all here.  The Efika was $100, seems like an accel based on it wouldn't have to be much more than about $300.  And it would even run super cool!  Just blue skying here..

[recidivist]

I think the only "legitimate" upgrades would be those  that provides a super-speed 68K cpu ,additional ram to the  limits of Amiga OS 3.x(perhaps 4.0 classic?),scandoubler/flickerfixer,and peripheral devices through usb and ethernet.

Once we no longer use the custom chipsets and their special features I believe it is no long an "Amiga",but rather "Amiga-like".

If the upgraded system can't use the software originally sold for Amiga then it must of necessity be considered something else,perhaps AmigaNG(new generation).

Still a 300 MHz 68030  might be fun!

[Karlos]

That would be HOT. I'd love to plug such a thing into my A500.  I always wondered why someone didn't just make a board which went into the 1200's accelerator slot, had a USB2 cable at the other end, and essentially made the 1200 a slave to the modern machine on the other end.  USB2's bandwidth exceeds anything happening inside the 1200, it could just see the other box as a really fast accelerator.

USB2 has a theoretical peak bandwidth of up to 480Mb/s (megabits/sec). In actual use you'll not see that sort of speed. My USB2.0 hard disk can pump data at around 30MB/s, which is certainly not faster than (for example) local memory speed on a good accelerator card.

Of course, the main problem with USB is not so much the theoretical bandwidth but the latency. USB is a serial, packet based protocol, pretty much designed for streaming data between devices. Latency there is not that big an issue, a few milliseconds point to point lag doesn't matter as long as the packets keep transferring at a continuous rate. That's not true for your A1200 trapdoor where you have dozens of independent (and interdependent) signals to worry about. Hardware interrupts and the like don't need to be held up by being encapsulated into a USB packet, sent down a wire to a host, which in turn will be interrupted to deal with it and so on (yes I realise USB has interrupt and control transfers but I don't think their timing requirements are necessarily suitable for dealing with stuff happening inside your miggy).

[blanning]

I think the only "legitimate" upgrades would be those  that provides a super-speed 68K cpu ,additional ram to the  limits of Amiga OS 3.x(perhaps 4.0 classic?),scandoubler/flickerfixer,and peripheral devices through usb and ethernet.

Once we no longer use the custom chipsets and their special features I believe it is no long an "Amiga",but rather "Amiga-like".

If the upgraded system can't use the software originally sold for Amiga then it must of necessity be considered something else,perhaps AmigaNG(new generation).

Still a 300 MHz 68030  might be fun!

Yeah, that's what I'm thinking.  Even a super speed 68000.  Choose the chip and clock speed.  The idea is that you could dial in the machine to whatever you want at the time.  If you want to play classic games, run the machine in super-speed 68000 mode for the UI, then hotkey to a menu to slow down the clock to 8mhz.  Choose 1.3 roms.  Later you can switch to a 68060 with 3.1 roms to do other things.

I'd still want to emulate the powerPC so I could run that software, or a newer version of the OS.  But not at the expense of actual 68k goodness.  You could even emulate some of the intel upgrades, like the golden gate 486 board, or that at-once 286 thing.

You could choose to use the amiga chipset sound hardware, or emulate various zorro 2 add-in sound cards.  I know there was a soundcard, can't remember which, that added 2 more channels with each one you put it.  Maybe the upgrade could provide virtual zorro slots for things like that, emulate 4 of those sound cards for 8 channels, but unify the output into a pair of RCAs.

What if instead of a 128mb or 256mb simm, you could plug in a 1gb simm.  Then you could partition the memory into multiple "machines".  Freeze one, then switch to another, sort of rebooting the 1200 when switching between them.

Basically, it's WinUAE and its hardware emulation on a 1200 accelerator upgrade board, or something to go in the 2000 CPU slot, with all the hardware interfaces.

brian

[lapeno]

Interesting ideas... It is good to have this brainstorming, maybe someone will use these ideas in the future.
I however still think that a normal 68k CPU board with 512MB SDRAM, a 030 and a 060 processor on it would be much more realistic. CPU could be selected on boot menu and in addition probably USB and Ethernet support could be added to the FPGA if there is space left in it. Next development step could be to replace the CPU with an FPGA based emulated 68k.

[lapeno]

I have just made some progress with planning and documentation reading. I am now concentrating on the memory controller part and found out that indeed it is not difficult to connect a 64bit wide SDRAM to the 32bit CPU bus. SDRAM modules use mask pins (DQM) with which you can easily mask out the upper or lower 32bit data based on address.
Here are some of my specs:
1. Using PC133 SO-DIMM memory modules (1 maybe two SO-DIMM slots)
2. Supported RAM modules 32MB,64MB,128MB,256MB,512MB which gives in theory 1GB max memory with two SO-DIMM slots. These modules are very cheap nowdays as second hand.
3. No burst mode due to lack of L2 cache. Next design will have 128K L2 cache.
4. 3.3v board using 68060 CPU
5. One old FPGA for 5v interfacing with Amiga, one new FPGA for memory controller and additional interfaces
6. possibility for future extension with 10/100 ethernet and USB 2.0 support on 2nd FPGA.
7. written in VHDL

[Crom00]

There are tons of options and FPGA's that can do quite a lot so you can have a CPU implemented in FPGA along with GFX, sound, etc. The only way something will get done is as a labor of love by an Amiga fanatic. Something like the minimig project where the sources BOM and PCB layout are released to the public. The Non recoverable engineering costs are a heavy load to bear for a small company.

[mumule]

I have just made some progress with planning and documentation reading. I am now concentrating on the memory controller part and found out that indeed it is not difficult to connect a 64bit wide SDRAM to the 32bit CPU bus. SDRAM modules use mask pins (DQM) with which you can easily mask out the upper or lower 32bit data based on address.

I was wondering when you notice that ;-)
Anyway, another way of doing this is to connect the 64 bits to one 32bit wide bus.
and switch the control signals. But, SDRAM is not the way to go anymore ...
To slow.

1. Using PC133 SO-DIMM memory modules (1 maybe two SO-DIMM slots)
2. Supported RAM modules 32MB,64MB,128MB,256MB,512MB which gives in theory 1GB max memory with two SO-DIMM slots. These modules are very cheap nowdays as second hand.
3. No burst mode due to lack of L2 cache. Next design will have 128K L2 cache.
4. 3.3v board using 68060 CPU
5. One old FPGA for 5v interfacing with Amiga, one new FPGA for memory controller and additional interfaces
6. possibility for future extension with 10/100 ethernet and USB 2.0 support on 2nd FPGA.
7. written in VHDL

to 1 & 2) If it has to be cheap, use on SO-DIMM. Also is much nicer to design cases it the long DIMMM Modules are not standing up, just as a reminder. Could be a nice flat board.
Personally, I would solder the RAM to the board. Less headaches with different versions of so-dimm modules/manufacturers

to 3) Careful ! You're missing something again. CPUs burst all the time, and you wanted to have an AGA in it to, right ? Yust make some calculations how much bandwidth you need and how much you get, using single accesses ...
(Entertain yourself with the 1600x1200 true color screen, and a 100 MHz CPU)

to 4) WHY ? You have a wonderfull big FPGA on board already. Why not use it ?

to 5) Big, expensive FPGA for CPU/MEMCtrl/VideoCtrl), the second for I/O is good. You actually run out of I/O pins very fast on those designs.

6.) Agree here, but put in on board already. People who design enclosures will love you ;-)

7.) Yes, however mixing Verilog/VHDL is not such a problem anymore.

Just my 0.000005 Cents ;-)

BTW, you just designed the Replay Board again, from http://www.fpgaarcade.com