Apollo Core 68080 is not only the fastest 68000 series CPU ever (as much as 500% faster than a 68060-66, 250% faster than a ColdFire MCF5474-266), it also is the most fully featured … Fully pipelined, superscalar, executes up to 4 instructions per clock cycle, two address calculation engines, two integer execution engines, market leading code density, optimal cache utilization, separate data and instruction caches, supporting concurrent fetch/read/write per clock cycle, automatic memory prefetching, memory stream detection, store buffer, branch prediction, and a fully pipelined double/extended FPU.
The m68k wasn’t the first processor I learned assembly for (that would be the 6502), but it was the one I wrote the most assembly for. I love that architecture. Years later, when I started having to deal with x86 for work, I was amazed at how…clunky…it was. x86_64 is a little bit better, but (partially for nostalgic reasons) nothing beats the m68k for me.
Motorola’s instruction sets are all extremely clean and consistent. I also learned 6502 assembly first and have an affinity for that architecture, but Motorola is great. Also, the 6809 is similar to the 68000.
In fact, Motorola produced the XLATE09 translation tool to convert 6809 assembly to 68000 assembly, which handled 85%+ of the work automatically.
For some reason I was expecting an open source implementation…
Maybe it’s the “When put in an FPGA” line and the “Downloads” link…
There are working 68k implementations in verilog with public source, though this one at least doesn’t have a proper FOSS license, and it’s not as good as the Apollo Core 68080…
Not FPGA (or a hardware CPU upgrade/replacement), but there is also the Cyclone 68000 and the UAE4ARM-RPI 68000 implementations for ARM.
I also wonder aloud … might this CPU be vulnerable to any SgxSpectre/MeltdownPrime/SpectrePrime/BranchScope/Meltdown/Spectre-like vulnerabilities due to the branch predictor?
Almost certainly susceptible to Spectre in one form or another. Probably not Meltdown.