Topic caught my eye as interest in the 6809 and floating-point. It is great to see a comparison of software packages. 6809 is one of my favorites. Know of any hardware floating-point for the 6809?
Not exactly software, but one late night I added 128-bit decimal floating-point support to my rf6809 core written in System Verilog as register ‘G’. add, sub, mul, div, cmp, load and store. Divide takes something like 2000 cycles. However, the number of instructions used is really low as it is all “hardware”.
The gcc6809 compiler supported/supports an external 68881 FP coprocessor, but how it’s connected is not clear. Let’s hope it’s more reliable than Innovative System’s FPE 68881 board for the Apple II from 1988. The only writeups I can find about that include more details about how often it crashes than the hardware itself.
If they were determined enough, and had a suitable power supply and cooling system, I’m sure someone tried to interface an Am9511 or Am9512 to a 6809 system.
Lastly, Micromega Corporation’s uM series of coprocessors were meant to be system-agnostic. While the website still exists, the product is unavailable as the company owner passed away a few years ago. I think the chips were based on PICs of some sort.