CommanderX16 and "What is my dream computer?"

[TomXP411]

3 hours ago, Sean said:

However, a 24 bit address space and floating point capability would actually let me do (in C, perhaps) much of what I'm messing around with in C# on my home computer.  I could have accomplished it with a little pain with a mid-80's PC or Mac, or maybe a IIgs, though with some workarounds.  I'm starting to understand why the concept of a so-called 3M computer was such the rage in the early 80's.

As an experiment, I'm working on a 16-bit instruction set with a pure orthogonal instruction set (ie: all instructions can use any addressing mode or register.) My long term goal would be to create an FPGA core and a BASIC and assembly programming environment for it. I eventually want to build this into a fantasy computer and turn that into an FPGA core.

I'm also pondering the idea of turning this into sort of a meta assembly langauge, where the orthogonal instructions get compiled into native instructions on 6502 and Z80 systems. It should be fairly straightforward, since nearly everything I'm thinking about can be mapped down to a small number of machine code steps. 



 

 

[hth313]

37 minutes ago, TomXP411 said:

As an experiment, I'm working on a 16-bit instruction set with a pure orthogonal instruction set (ie: all instructions can use any addressing mode or register.) My long term goal would be to create an FPGA core and a BASIC and assembly programming environment for it. I eventually want to build this into a fantasy computer and turn that into an FPGA core.



 

Did not the PDP-11 do this 50 years ago?

[TomXP411]

23 minutes ago, hth313 said:

Did not the PDP-11 do this 50 years ago?

Probably. But I don't want a PDP-11... I want MY dream computer. ? (I mean, I have a PiDP-11... and have not done anything with it other than watch the blinkies.)

 

 

[Kalvan]

I'll have to describe my personal dream computer in the general retro-computing forum but I hope that it isn't too much of a hijack to post a few comments to second guess Mr. Murray's chip selection a bit.

Personally, the way Mr. Murray and his team, in terms of remaining true to the vision and ideals they have set themselves up for the Commander X16, have done as perfect a job as could be expected without commissioning masks for VERA (and separating it from the Geometry Synthesis/PCM sound engine) with Samsung, TSMC, or Global Foundries, or gone to some FPGA to ASIC pathway specialist.  Still, I would have chosen as my CPU either the Hudson Hu6280 or the Nintendo SA-1, assuming I could find some means of assuring a steady supply that did not involve cannibalizing them from classic consoles and game cartridges.  I know, they're surface-mount chips, and IBM's patent on it didn't expire until 1987, not to mention the hassle integrating them on a mostly DIP socket motherboard, but you get demultiplexed address and data pins and full 16-bit data busses, with no need for external bank switching.  In fact, you get three, at least one of them full-duplex if my interpretations of the pinouts are accurate, not to mention the Hu6280's six wavetable sound channels and mass-move instructions or the SA-1's two extra registers and hardware multiply and divide instructions in 16-bit mode.  I would have also possibly added in a Super Nintendo cartridge add-in DSP or two (Their cores date to 1982), with the cover story being that this computer is released in 1987-88, with Hi-Toro Labs having been bought out by someone else, as competition for the Amiga, Sharp X68000, Acorn Archimedes, MSX 3/Turbo R machines, and VGA and Ad-Lib/Sound Blaster era PC Clones.

But that's just me.  Personally, in Mr. Murray's shoes, and without his example, my first effort at this sort of thing would have been much kludgier.

[xanthrou]

On 1/21/2021 at 11:26 PM, picosecond said:

2/10: OFF THE SHELF COMPONENTS



Gray market YM2151/3012 do not meet the criteria.  I am quite certain others will disagree with this opinion.

Yeah, but David commented that he wouldn't mind YM2151, so 10/10.

[BruceMcF]

16 minutes ago, xanthrou said:

Yeah, but David commented that he wouldn't mind YM2151, so 10/10.

While it's subjective how BIG the steps should be on an ordinal scale, you do have to HAVE steps for situations that are "clearly better" and "clearly worse". And you aren't giving it headroom ... while he said he wouldn't mind a Yamaha FM chip, in the context of the off-the-shelf criteria, a YM2151 in production would have to be superior to a YM2151 from pulls, and then if "wouldn't mind" sounds like less than ideal, if an in production YM2151 is part of a 9/10 score, then a grey market one would be 8/10 at most.

In tail room, you'd have to leave room for the processor, I/O chips, decode circuitry being new-old, grey market or CPLD/FPGA simulated stock, so I can't see enough tail room on the scoring if the score is set any lower than 5/10.

Where that score is between 5/10 and 8/10 ... I dunno, I gave it a 7/10 as a C, but I wouldn't don boxing gloves to defend that spot within that range.