A few points:

1) The RaspberryPi has SPI, I2C and serial easily available on the GPIO connector.

2) The RaspberryPi is cheaper than the "relatively cheap" graphics option that you mention; it also offers higher resolution and much higher communications bandwidth.

To me, the RaspberryPi is a cheap building-block for all sorts of things - including providing services to 6502-based systems.

I'll be getting my RaspberryPi later this week; then I'll see how good it actually is

Rpi as a graphics or audio peripheral would be a nice application!
Even mass storage, given the SD card. Oh, and networking too...
I like this. Especially coupled with a 100MHz FPGA-based 6502!

I know that you all like this forum to be neat, but i never thought there would be such trouble because of the topic. I started this topic because i wanted to see what you thought about the device and its uses, and not the people on the r-pi forum, i am already registered there and posting...
One other thing is that the pi is the only small computer that inspires people to learn to program, and it keeps the spirit of the 80' era and of the bbc micro, which is a 6502 machine.

If we want to have mixed projects with the 6502 and pi, the pi shall be first introduced in the community(which is what i was trying to do).
As for my comment that the 6502 was not fast enough for video, i meant that the chip doesn't have enough mips for loading video from a video camera. I started a topic about how to connect a video camera to a 6502 device, but i got no reply. The pi is less expensive than an fpga board, so ill go for the pi do to the video job.

Hi BDD
Hope you don't mind a quick comment on this:

Graphics aside, the Pi is about as open as any Linux machine, and indeed can be programmed at bare-metal instead if one wishes. On the graphics front, one can use the X drivers, or OpenGL ES - both are well documented APIs. The 3D demos are quite impressive. I don't know about audio or about video streaming although obviously both can be done. Audio is certainly available using standard linux mechanisms.

The GPIOs are accessible from python or from command line, from C or of course from bare metal.

Cheers
Ed

The area of microprocessor system design and implementation is enormous.

The nice thing about this forum is that it focuses mostly on a single MPU, that being the 6502. And it focuses on it outside of the systems built around it. It's lower level. This is not an Atari, C64, or Apple forum per se (though there is inevitable overlap).

A large amount of what is discussed here is applicable to other processors, and there really isn't, as far as I know, a really comparable forum for other microprocessors. I don't think there's a Z80 forum working at the low, signal and board level that this forums works.

Arguably, there really doesn't need to be. Most of what folks talk about here is really applicable to a someone trying to wire wrap a Z80 based board, and if someone came here with assembly questions, schematic questions, signal questions about such a project, they'd probably get a warm reception, since there's so much overlap and the project would interest even folks here.

But deciding to focus on a single MPU I think is valuable as it does keep the board from getting distracted.

In the end, particularly today, the general purpose MPUs of today are just that -- general purpose, and usable for pretty much anything, anywhere. The combination of cost, available components, tool chains, and lore based "trivia" knowledge learned over long experience have more impact on MPU choice than most anything else.

I come here partly out of nostalgia, partly out of experience, but mostly because I wanted to work with something I could get my head around. The 6502 fits that pretty well.

Soldering an SOC to a board, applying power, and booting Linux, that's not what I was looking for. I can program linux all day long. One SSH session looks like any other.

I looked at what I wanted to do, and rather than buying a development board, I wrote a simulator, as for what I'm doing there's no difference between a simulator and a board with a serial terminal.

The Pi is interesting because it's cheap. Other than that, it's just another Linux box. If offers the same experience as any other SSH prompt. I can program Python on anything.

The concept of a using the Pi as a cheap "universal peripheral" for a 6502 project is...interesting... but, then, what's the point? Why not just use the PI for the rest of it?

That said, it's interesting in the same way that Garth's work bench computer is interesting, he uses that as his testing Swiss Army Knife. You can play 6502 and then use the Pi for "everything" else. That's compelling, again, depending on what you're looking to do. In my case, I find the GUI parts of my simulator annoying -- because it's not what I want to spend my time on, I don't find it interesting. Having the Pi as a universal adapter may make the 6502 (or anything else) more interesting because now, for the short term, they don't have to worry about interfacing Network, USB, or video to the 6502 -- you can just proxy that via the Pi -- for $35 and some wire.

So, I think that would be an interesting project -- coming up with a protocol or design spec for interfacing a Pi via either parallel and/or serial, and then how to write "drivers" on the Pi to proxy the devices. It seems a "waste" to use a Pi as a mere super duper VIA chip, but for $35, you CAN use it for that, for a personal project (I don't think you'd ship anything like that). Consider the various 6502 video projects using micro controllers to drive VGA displays. Well, here's something that will fit that bill, and likely be cheaper.

Getting a Pi to DMA video pages out of a running 6502 SBC -- that's interesting isn't it? A live 6502 with a simulated ANTIC or VIC chip? That's interesting also.

I mean, I'm simulating the 6502 in a Java JVM. I'm almost as far away as you can get from the 40 pin DIP, but it's still interesting. Having a peripheral processor running 300 times faster than the Host chip is still interesting.

Other than that, the Pi, as a computer system in its own right, is simply too high level to interest me on this forum.

This is why, if the Pi lacked the GPIO or was not cheap (or was not open), it wouldn't fit at all.

JGH has written a Tube-over-serial protocol implementation (and spec) which offers the BBC micro's idea of a host processor with services and a second processor as a compute element. I haven't the time, but in principle find it interesting to use that protocol to offer high-level services over a serial channel, using an existing 6502-compatible API. At present the host side is in BBC BASIC, so the first step would be to re-implement in python or C.

In my case the 6502 computer using the Pi host could be an ARM-emulated 6502 on the cheap Discovery board... (No matter that it runs somewhat slower than a 6502 emulated directly on the Pi.)

Cheers
Ed

One other issue with the pi is that it has only 8 gpio pins. Some time ago i was thinking how it would be possible to make an interface that could be used to control 65xx peripherals. Here 6502 could manage and decode signals from the pi, and then have an unlimited number of peripherals, from more gpio using the 6522, to other kind of i/o. And if somebody is lucky enough to have a sid chip, a synthesizer could be made...

I think the general tactic when short of GPIO is an SPI expander (i2c would also be possible, but presumably slower?)
Cheers
Ed

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

Quick reply BDD: different people engage at different levels. Some like to solder things together, some like to program, some like FPGAs. Yes, the Pi is a solid piece of hardware, but the interfaces at the edge of that hardware are useful and well described. To the software on the Pi, it is open. In both cases, the skills learnt in using the Pi are transferable - and the Pi is all about education.
Cheers
Ed

Have previously had an unsatisfactory trial with the 'BeebIt' pair of bbc micro emulators on the Pi, today I ran up the xbeeb emulator. (There's a helpful binary available, but it needs tweaked sources, so I re-tweaked the source to make a comparable build of my own. I wanted to run on the raspbian distribution which targets a closer approximation of the ARM on the pi)

Result: xbeeb does run. However, it runs fast (*), so '*fx11' is needed to stop key repeat from making typing impossible. Also the TIME function in BASIC runs fast. I wasn't able to run the included game - something not right with the emulated connection to the local filesystem.

So, again not totally satisfactory.

(*) Maybe running about 8MHz, so 4x normal Beeb speed. It could go faster, but then it's impossible to type the '*fx11'. One fix would be to nobble the VIA division constant in the ROM which produces the realtime interrupt. Another would be to rejig the emulator to slow itself down if it's overtaking reality - but for some purposes a faster Beeb is a good thing.

Would a I2C GPIO expander work? I suppose anything with enough wires is workable.

http://www.mouser.com/Search/Refine.asp ... O+expander

I was thinking about building a small robot, but i would like to have a 6502 with it too, and before you say that the pi can emulate it faster that it can work, there is another reason. The 6502 would be a watchdog device that could reboot the pi in the case of a crash, and take the control of the robot, and handle all the i/o operations, to avoid the expander mess. But i am wondering about how could i get raw data from a camera using the pi, and it seems the pi doesn't have an audio in so the 6502 should handle that too.

Weeeeeeeeeee!

I got the pi finally, actually i got two, one for me and one for my friend.

Excellent!