The 6809 was almost certainly the ultimate 8-bit processor design since it had many 16-bit features as well as advanced addressing modes, stack handling and advanced instructions including a multiply.
The crown of "Best 8-Bit CPU" should probably go to the little known 6309 CPU, an advanced variant of the 6809 made by Hitachi under license from Motorola (https://en.wikipedia.org/wiki/Hitachi_6309). It booted up fully compatible with the 6809 but could be switched to a 'native mode' where it would execute 6809 instructions 30% faster, have more registers added and a significant number of additional instructions including divide. Being fabbed in CMOS, it could also be clocked much higher than the 6809's 2 Mhz, up to 3.58 Mhz.
Anyone wanting to create "The Ultimate 8-Bit Microcomputer" from early 80s components would want to start with a 6309 as the CPU and perhaps marry it to the 9958 video display processor used in 8-bit MSX2 computers which was probably the most advanced VDP of the era since the 6847 VDP Motorola designed for the 6809 was extremely basic with no sprites or extended palette features. For sound the best 8-bit era chip would be one of the Yamaha FM chips commonly used in MSX2 machines. Unfortunately, no manufacturer ever married these components into a world-beating Ultimate 8-Bit. It would have been an amazing machine for the time and certainly held the position as top hobbyist lust-bait - at least until the arrival of the 68000-based Amiga.
I don't know about the best, but to me it is the most beautiful!
And with well conceived video hardware, the 6809 could do a lot!
The arcade game DEFENDER is a single, pixel pushing 6809 drawing to a single frame buffer. Notably, that hardware included a signal when the CRT raster hit the middle of the screen!
This made a single bitmap frame buffer function as a double buffered display, without the cost of having to clear an entire screen in between frames.
You can skip through that to find some of the real action. Quite impressive!
Notably, a 6800 does the sound. One sound at a time generated parametically.
There have been some experiments to drop a 6809 into other 8 bit systems. One was created for the Atari, but no real software pushing limits was ever created.
I would think one of the better transplants would be the Apple 2. It has a video system that does not interrupt the CPU. That would operate much like the Williams hardware, minus the handy screen addressing.
The 6809 is my favorite 8 bit CPU.
While I have not programmed a 6309, I would enjoy it immensely.
Maybe I can stick one onto a card in my Apple one day.
> I don't know about the best, but to me it is the most beautiful!
Since I was a relatively impoverished college student and had no prior experience with computers when I got my first, a 4k 6809-based Tandy Color Computer, I learned Coco BASIC and then 6809 assembly language (using Radio Shack's ROM cartridge-based Assembler/Editor), I had no inkling how spoiled I was. The Coco's Extended Color BASIC was quite advanced for the time (as compared to the stock ROM BASIC's on comparable micros) and the 6809's ISA was leagues ahead of the 6502 and Z80. I had no idea until much later (post 8-bit era) how relatively primitive the ISAs were on the 6809's 8-bit peers. Since it was all I knew, indexed and program-counter relative addressing just seemed like the obvious natural way of things.
I also didn't understand what was meant by "orthogonal instruction set" when I read the term in Osborne's 6809 book. I just thought every instruction and addressing mode for each register and stack pointer would always have a complete set of its logically-implied counterparts in all CPUs! :-) Oblivious to how fortunate I was in these ways left me feeling especially jealous of the dedicated graphics and sound hardware of the Atari and Commodore 8-bits.
Learning on the 6809's advanced architecture along with the multi-tasking, multi-user, UNIX-like OS-9 operating it enabled, left me uniquely well-prepared for the future of computing in ways other popular 8-bit micros couldn't and I never even appreciated it at the time.
Yeah, the Moto VDA was crappy! 256x192 with some odd color and semi-graphics was not exactly brilliant.
The CoCo 3 ended up being a pretty great machine, but too late.
I had a bit of a blast programming that one briefly in one byte per pixel mode! On NTSC models, one can set the 640 pixel, 4 color mode or 320 pixel, 16 color modes and use a composite display to get basically 256 artifact colors!
I had only a cassette and limited time. All I managed to do was some nice fractal plots, and the like in assembly language.
Today, we have a fair number of good productions on that machine, though surprisingly few make use of the excellent composite modes.
Yeah, the whole setup, if one had it, really was a great education.
The crown of "Best 8-Bit CPU" should probably go to the little known 6309 CPU, an advanced variant of the 6809 made by Hitachi under license from Motorola (https://en.wikipedia.org/wiki/Hitachi_6309). It booted up fully compatible with the 6809 but could be switched to a 'native mode' where it would execute 6809 instructions 30% faster, have more registers added and a significant number of additional instructions including divide. Being fabbed in CMOS, it could also be clocked much higher than the 6809's 2 Mhz, up to 3.58 Mhz.
Anyone wanting to create "The Ultimate 8-Bit Microcomputer" from early 80s components would want to start with a 6309 as the CPU and perhaps marry it to the 9958 video display processor used in 8-bit MSX2 computers which was probably the most advanced VDP of the era since the 6847 VDP Motorola designed for the 6809 was extremely basic with no sprites or extended palette features. For sound the best 8-bit era chip would be one of the Yamaha FM chips commonly used in MSX2 machines. Unfortunately, no manufacturer ever married these components into a world-beating Ultimate 8-Bit. It would have been an amazing machine for the time and certainly held the position as top hobbyist lust-bait - at least until the arrival of the 68000-based Amiga.
And with well conceived video hardware, the 6809 could do a lot!
The arcade game DEFENDER is a single, pixel pushing 6809 drawing to a single frame buffer. Notably, that hardware included a signal when the CRT raster hit the middle of the screen!
This made a single bitmap frame buffer function as a double buffered display, without the cost of having to clear an entire screen in between frames.
Very efficient.
https://m.youtube.com/watch?v=wAKxa5C9jHY&t=3122s&pp...
You can skip through that to find some of the real action. Quite impressive!
Notably, a 6800 does the sound. One sound at a time generated parametically.
There have been some experiments to drop a 6809 into other 8 bit systems. One was created for the Atari, but no real software pushing limits was ever created.
I would think one of the better transplants would be the Apple 2. It has a video system that does not interrupt the CPU. That would operate much like the Williams hardware, minus the handy screen addressing.
The 6809 is my favorite 8 bit CPU.
While I have not programmed a 6309, I would enjoy it immensely.
Maybe I can stick one onto a card in my Apple one day.
Since I was a relatively impoverished college student and had no prior experience with computers when I got my first, a 4k 6809-based Tandy Color Computer, I learned Coco BASIC and then 6809 assembly language (using Radio Shack's ROM cartridge-based Assembler/Editor), I had no inkling how spoiled I was. The Coco's Extended Color BASIC was quite advanced for the time (as compared to the stock ROM BASIC's on comparable micros) and the 6809's ISA was leagues ahead of the 6502 and Z80. I had no idea until much later (post 8-bit era) how relatively primitive the ISAs were on the 6809's 8-bit peers. Since it was all I knew, indexed and program-counter relative addressing just seemed like the obvious natural way of things.
I also didn't understand what was meant by "orthogonal instruction set" when I read the term in Osborne's 6809 book. I just thought every instruction and addressing mode for each register and stack pointer would always have a complete set of its logically-implied counterparts in all CPUs! :-) Oblivious to how fortunate I was in these ways left me feeling especially jealous of the dedicated graphics and sound hardware of the Atari and Commodore 8-bits.
Learning on the 6809's advanced architecture along with the multi-tasking, multi-user, UNIX-like OS-9 operating it enabled, left me uniquely well-prepared for the future of computing in ways other popular 8-bit micros couldn't and I never even appreciated it at the time.
The CoCo 3 ended up being a pretty great machine, but too late.
I had a bit of a blast programming that one briefly in one byte per pixel mode! On NTSC models, one can set the 640 pixel, 4 color mode or 320 pixel, 16 color modes and use a composite display to get basically 256 artifact colors!
I had only a cassette and limited time. All I managed to do was some nice fractal plots, and the like in assembly language.
Today, we have a fair number of good productions on that machine, though surprisingly few make use of the excellent composite modes.
Yeah, the whole setup, if one had it, really was a great education.