Despite how simple and limited the 6502 is, it's always interesting to see this almost 50-year-old architecture being stretched to new limits. 6502 cores are still found in some SoCs aimed at the ultra-low-cost high-volume market.
> 6502 cores are still found in some SoCs aimed at the ultra-low-cost high-volume market.
What's the reason behind that? Probably just keeping existing, working designs and thus saving R&D costs? Sounds like beating a 10 cent risc-v core would be hard.
Edit: When I first read your comment I had to chuckle a bit, because I got the image of a multicore SoC with 6502 6502 cores. That would be a fun project for my FPGA hmmmmmmmm ...
10c per part is incredibly high depending on how low-cost we're talking.
If your chip needs very basic processing, and you're going to sell it for less than a cent per chip, you need your core to be basically free.
Sure RISC-V is very inexpensive, but for parts like these risc is massive overkill and as a modern core, you'll be spending a lot of die area on a core that should really be an ASIC, but designing an ASIC is expensive so you can just slap a 6502 in that bad boy and call it a day.
On the other end of the scale, you'd be surprised to learn that many chips you don't ever interface with as a user contain even ARM cores. These are usually ASICs in high-end products.
16 registers * 32 bits * 4 transistors / cell = 2k transistors just for storing the bits alone, and that's not including the decoding and read/write circuitry.
There is a ton of code out there for pretty much any task you want.
I have the 16 bit successor, the 65816, on a variable speed expansion card in my Apple 2. I run it in 8 bit mode most of the time mostly because it works and most library code is 8 bit.
At those speeds, the chip is fast! Especially when one considers the simple model: ram clocked 1:1 with the cpu. In my case, I can run code over the 1Mhz bus, essentially turning most multi cycle ops into a single bus cycle per memory fetch.
Or, the card has a megabyte of RAM and that runs at CPU speed. (0.15 to 16 Mhz) When doing that, it is fast enough to run large programs, written in higher level languages, at a useful speed. Of course Assembly language is crazy fast.
This post has been hanging out with no comments for ages. I was like, I'll skim it later.
> It's almost burying the lede to announce that, though, because the real meat in this entry is how the Commodore 64 manages to emulate a very different 6502-based system. That piece is "6o6," for "6502-on-6502," and is a full virtualized software NMOS 6502 CPU that runs on a 6502 CPU — which I've open-sourced too. It has full control of guest code execution, including trapping undocumented and jam opcodes, and completely abstracts all memory access, making it possible to remap addresses, intercept illegal reads or writes, or even run entirely from virtual memory. On top of that, it's complete enough to not only pass a full functional test but also virtualize itself virtualizing itself:
This is incredible! Not just from a 6502 perspective, but from an everything perspective! Amazing work.
> There is a catch (rather: a couple) of course: It runs extremely slowly and it needs a RAM Expansion Unit (REU), as there is no chance to fit it all into just 64KiB.
This is amazing. It brings back memories of the first time I taught myself 6502 assembly, there was a book called “The Visual Computer” that came with an emulator on a floppy disk. It was very eye opening! I found a PDF of the book [1] but not sure if the software on the floppy survives anywhere.
Not sure about the C64 one but I've seen the Apple ][ one on the software archive sites before sites before. It's The Visible Computer, a reference to these educational models:
This one hell of a bucket list you must have. Respect.
reply