FPGA vs Real, what is the tradeoff?

[Falken]

20 hours ago, StephenHorn said:

Yeah, that's probably my fault. ? I saw mentions of other new 6502-based systems, and in my head I immediately drew analogies to this one.

There sseems to be no problem designing a 6502 computer that simply computes but making one that has simple graphics and sound without resorting to FPGAs seems to be next to impossible.

[m00dawg]

4 hours ago, Falken said:

There sseems to be no problem designing a 6502 computer that simply computes but making one that has simple graphics and sound without resorting to FPGAs seems to be next to impossible.

I don't think it's impossible, just very expensive and likely pretty large due to all the parts required.

[SlithyMatt]

5 hours ago, Falken said:

There sseems to be no problem designing a 6502 computer that simply computes but making one that has simple graphics and sound without resorting to FPGAs seems to be next to impossible.

You basically have two other options, both orders of magnitude more expensive: ASIC (but that expense can be mitigated by volume) or discrete logic (which will not only be much more expensive, but require a large expansion card, taking up one of the slots by necessity)

[Cyber]

On 1/9/2021 at 4:52 PM, Falken said:

There sseems to be no problem designing a 6502 computer that simply computes but making one that has simple graphics and sound without resorting to FPGAs seems to be next to impossible.

I think it's possible, but your video and sound solutions would be expensive (like was noted in previous answers) or very primitive.

There are off-the-shelf ready video and audio soultions on the market with reasonable prices, but usually they either have low specs and primitive features or they have some limitations. Some of them are mentioned in your Simple GPU thread.

Also it is possible to build very simple and not expensive video card yourself. Watch Ben Eaters video about world's worst video card (and its part 2, part 3 and part 4). Ben showed just a proof of concept, by I think it can be improved to run with tolerable performance without making it expensive.

Also I think a similar sound card can be done. I know many people wish Ben to build word's worst sound card. I hope he will. )

[Falken]

On 1/10/2021 at 6:57 PM, Cyber said:

Watch Ben Eaters video about world's worst video card (and its part 2, part 3 and part 4).

I already did and was really impressed. It reminded me of the early EGA and Hercules cards I used. I wish I could build something like that. I am only now slowly coming back to a more low-level approach to computing and try to program a little ATMega gadget. So I might sometimes ask questions that are very low level or are ignorant of certain diskussions that already happened.

 

On 1/10/2021 at 6:57 PM, Cyber said:

Ben showed just a proof of concept, by I think it can be improved to run with tolerable performance without making it expensive.

"Expensive" is a relative term, depending on the buyer. Most of us here are probably adults with some expendable income, so for a machine we really want a couple of a 100€ is probably affordable. For something with a broader appeal you'd probably need a price < 100 € to attract parents to just get "something educational" for their kids. Ben's card alone as individual components is 100$ . The Arduino managed to get into that particular nieche because it was < 50€ on average I think.



 

[Travis Bryant moore]

As for Feature creep i think FPGA is just a modernization feature. And if you can play the odd Weird Al's it's all about the Pentiums,  so what. I mean I think the goal is to do an 8 bit machine and make it fun to use. Retro machines may just be more for a hobby rather than the next new retro gaming console. Could you imagine getting a contract to make a million of these things for Mattel or some toy company and include software and what not for a price point under 100 dollars? A retro computer could do a lot of things so that may not be an easy task to add all bells and whistle so to speak.

[TomXP411]

The tradeoff is that FPGAs are rarely perfect recreations. As often as the 6502 processor and the Commodore 64 have been replicated in FPGA (there are three distinct projects I'm aware of), the emulation is still imperfect. In fact, the software emulation on VICE passes more test cases in the VICE test suite than the Ultimate 64, Turbo Chameleon, or MiSTer C64 cores. 

 

[paulscottrobson]

On 1/3/2021 at 10:19 PM, StephenHorn said:

I think it's fair to argue that the drawbacks to FPGAs mostly have to do with the "purity" of a project.



But I think another consideration is the constraints of the project. I mean, if you're not going to use real hardware and live within its limitations, why bother pretending to live in a world of real hardware? Increase the clock speed, change the instruction set, add native 4K resolution with 3D graphics... creep those features all the way to a modern X86-64. In fact, why not skip all this hardware nonsense and stick to making a new microconsole-of-the-week or framework-of-the-month? It's all virtual, anyways.



You can't do that so easily when you're working with actual hardware components. And that, too, is part of what makes accomplishing a project like this interesting.

Well, except now the only thing that is actual CPU is the core itself. The VERA fpga fails entirely on that criteria

[Elektron72]

6 hours ago, paulscottrobson said:

Well, except now the only thing that is actual CPU is the core itself. The VERA fpga fails entirely on that criteria

There are multiple examples of real chips being used in the X16. Examples include:

• 
  The 65C02 CPU
  


• 
  Two 65C22 VIAs
  


• 
  The YM2151 FM synthesis chip
  


• 
  RAM
  


• 
  ROM
  

The X16 would still be a functional computer without the VERA. Furthermore, the VERA does not have unlimited capabilities. Posts from those working on the hardware have indicated that the current design (which integrates video, PSG/PCM audio, and a simple SD interface) almost reaches the limits of what the current FPGA core can do.

[StephenHorn]

9 hours ago, paulscottrobson said:

Well, except now the only thing that is actual CPU is the core itself. The VERA fpga fails entirely on that criteria

I'd suggest you find a viable alternative, but you'd be years too late on that front. A number of video choices were evaluated and rejected, and in the end an FPGA of some sort was the only real option here.

And part of me does lament the integration of the SD interface, as well as the PSG and PCM audio, into the same FPGA, exactly because those decisions run counter to the "purity" goal of the project. However, once the choice was made to go with an FPGA for video, it's easy to see how the scope of that FPGA's responsibilities grew, not only to keep the final cost of the X16p lower, but especially if the long-term plan would be to eventually release an X16e that might be reduced all the way to something like a single FPGA. I'm sure it also helped to keep the design of the X16's motherboard a little simpler.

But as Elektron points out, there are still plenty of components that are good, old-fashioned silicon. The VERA has not become the be-all, end-all of the X16p.