Commander X16

29 minutes ago, picosecond said:

I have never believed these are connected.  I think it is unfortunate that 8BG has been promoting this notion.



The only thing that makes off the shelf parts understandable is their documentation.  Without docs how could anyone design with them?  Even good docs stop at some level of abstraction.  For example, YM2151 docs describe nothing about its microarchitecture, which is need to really understand how it works.  I would argue that properly documented highly integrated designs can be more understandable than their off the shelf cousins.  Phase 1 X16 and Phase 3 X16 are equal complexity and equally understandable.  The packaging differences are superficial.



Raspberry Pi has no architectural commonalities with phase 3 X18/X8.  The only superficial thing they have in common is a high level of integration.



 



If people prefer the cool appearance of big PCBs with lots of chips, I have no argument with that.  I think they look cool too.  I just reject this idea that knowing this chip is the CPU and this chip does graphics imparts any meaningful knowledge of the computer's operation.

I agree for the most part. The one place I think the big board with lots of chips wins is when trying to explain or teach it to people who have no idea. A concrete chip that can be described verbally is going to win out over a datasheet from a concrete vs abstraction perspective.

2 minutes ago, Scott Robison said:

I agree for the most part. The one place I think the big board with lots of chips wins is when trying to explain or teach it to people who have no idea. A concrete chip that can be described verbally is going to win out over a datasheet from a concrete vs abstraction perspective.

Also for troubleshooting / maintenance, if a chip get fried it's easier to replace and certainly cheaper. I agree it's not going to happen a lot but who knows, playing with the hardware (expansion cards) can probably cause some issues ?‍♂️

6 minutes ago, Scott Robison said:

I agree for the most part. The one place I think the big board with lots of chips wins is when trying to explain or teach it to people who have no idea. A concrete chip that can be described verbally is going to win out over a datasheet from a concrete vs abstraction perspective.

I get your point, but I would call what you are describing superficial understanding, not meaningful understanding.

The idea I am ranting against is that discrete implementations are necessary or superior for deep understanding.  That does not match my experience.  As an artistic choice, great.  As a pedagogical choice, not so much.

46 minutes ago, picosecond said:

I have never believed these are connected.  I think it is unfortunate that 8BG has been promoting this notion.



The only thing that makes off the shelf parts understandable is their documentation.  Without docs how could anyone design with them?  Even good docs stop at some level of abstraction.  For example, YM2151 docs describe nothing about its microarchitecture, which is need to really understand how it works.  I would argue that properly documented highly integrated designs can be more understandable than their off the shelf cousins.  Phase 1 X16 and Phase 3 X16 are equal complexity and equally understandable.  The packaging differences are superficial.



Raspberry Pi has no architectural commonalities with phase 3 X18/X8.  The only superficial thing they have in common is a high level of integration.



 



If people prefer the cool appearance of big PCBs with lots of chips, I have no argument with that.  I think they look cool too.  I just reject this idea that knowing this chip is the CPU and this chip does graphics imparts any meaningful knowledge of the computer's operation.

Off the shelf parts (not programmable chips) have clearly defined functional boundaries and electrical interfaces.  This means each part is responsible for something.  The interaction between the chips is meaningful, and the circuit can be modified, customized, repaired, etc.  This is mostly sacrificed when all the functionality is simulated in a programmable chip.  The organization of the hardware can and does inform understanding of how the computer works, and simpler and more transparent design makes it easier to understand.  If none of this matters, then yes by all means let's just forget about phase 1 and 2 and go straight to FPGAs.  I'm not going to suggest that this makes the product pointless, but it's not at all what 8BG presented.

The Raspberry Pi is as you say highly integrated, and its operating system is Linux which is not ideal for personal mastery because of its size and complexity.  This can often be too much for the beginner or casual hobbyist.  The Raspberry Pi is a great product, don't get me wrong, but conceptually totally different from an X16.

36 minutes ago, picosecond said:

The idea I am ranting against is that discrete implementations are necessary or superior for deep understanding.  That does not match my experience.  As an artistic choice, great.  As a pedagogical choice, not so much.

What group of students are we talking about, specifically? For a young kid, having individual components that can be visually segregated definitely helps. Once you hit a certain level of understanding abstraction, the physical representation is less important, but foundationally I think it's very useful.

2 hours ago, BruceMcF said:

65816, 24bit address bus, directly addressable video RAM, system ROM in Bank 0, User ROM is Bank 1, dedicated dual serial chip, separate microcontroller to access the SD card, PSG with (a limited amount of) hardware ADSR ... there's been a substantial amount of feature pruning as well.



Even later, the Vera UART functions which had been added were stripped out to allow an increase in the addressed registers from eight to thirty-two.



And of course, features that were not pinned down could be seen as feature creep by someone who imagined something simpler and a clamp on features by someone who imagined something more ambitious.

Yeah, no dispute here...the waters have definitely been muddied as this project has proceeded, which is 100% like every single other engineering endeavor I've been a part of. Depends on which angle you're looking at it, I suppose.

15 minutes ago, Carl Gundel said:

Off the shelf parts (not programmable chips) have clearly defined functional boundaries and electrical interfaces.  This means each part is responsible for something.  The interaction between the chips is meaningful, and the circuit can be modified, customized, repaired, etc.  This is mostly sacrificed when all the functionality is simulated in a programmable chip.

It feels like you are conflating microcontrollers and FPGAs.

Logic implemented in an FPGA has clearly defined functional boundaries and electrical interfaces.  Each part is responsible for something. Interaction between functional units are meaningful.  Given the necessary design collaterals the circuit can be modified or customized.

Some compiled C++ program which simulates a system is completely different of course.  But this has nothing to do with an FPGA implementation like CX8.

Sure, the 6502 core in CX8 won't have exactly the same microarch as one you bought from WDC.  The building blocks are different.  But they are not all that different.  The FPGA implementation is "just" a bunch of gates, flops and RAMs all wired together.

32 minutes ago, Carl Gundel said:

The Raspberry Pi is as you say highly integrated, and its operating system is Linux which is not ideal for personal mastery because of its size and complexity.  This can often be too much for the beginner or casual hobbyist.  The Raspberry Pi is a great product, don't get me wrong, but conceptually totally different from an X16.

I don't know what prompted this comment.  I never advocated for the Raspberry Pi here and specifically said it is a completely different architecture.

44 minutes ago, picosecond said:

I get your point, but I would call what you are describing superficial understanding, not meaningful understanding.



The idea I am ranting against is that discrete implementations are necessary or superior for deep understanding.  That does not match my experience.  As an artistic choice, great.  As a pedagogical choice, not so much.

I agree, certainly the non-discrete components can be understood. From my perspective, I think of the big board with chips being easier to try to explain to my middle school computer class because there is a more concrete divide between chips rather than an all in one FPGA or SoC solution.

28 minutes ago, picosecond said:

I don't know what prompted this comment.  I never advocated for the Raspberry Pi here and specifically said it is a completely different architecture.

That's true, sorry.  I think I was simply further making my case against the RPi vs the X16, which isn't clearly relevant to what you wrote.

4 hours ago, picosecond said:

I just reject this idea that knowing this chip is the CPU and this chip does graphics imparts any meaningful knowledge of the computer's operation.

I disagree.

I've watched a bazillion repair videos, many of them arcade machine repairs. Many times, the failure mode tells you which of the chips is faulty - like if the even pixel rows of tiles are garbage but the odd rows are correct, that means probably the LSB of some logic chip has failed or is stuck hi/low. You can see how the computer does its machinations when the traces between the CPU and the RAM lead through decode logic.  Sure, on the INSIDE, this makes no difference whatsoever. STA $370c is going to make whatever's in the accumulator go to that address, regardless whether the entire transaction happens beneath an epoxy blob or if it heads out across copper traces and gets manipulated by a bunch of logic chips, splitting the bits up between a bunch of 1-bit DRAMS. You can capture these traces with an O-Scope or logic analyzer or whatever and see what's actually taking place. From the inside, the CPU says "STA $9F41" and you don't hear the FM playback change according to the value you just wrote - what now? What's broken? Not your code. Not the architecture. A real chip failed to place a value into another real chip or that other real chip failed to produce the expected output on a pin somewhere, or maybe it did and that signal didn't make it to the speaker jack.....

That's the difference. One's a "cyberspace" view of the architecture, and one's a "realspace" view of the architecture. You can say that you know the memory map of a computer system; that you know what the function of each and every memory-mapped register is. You can know exactly what the procedures are for communicating with devices at those addresses (or behind those addresses for things like VIA-attached stuff)... but that doesn't mean you know HOW the computer works. It just means you know how the computer WORKS.



See the difference?