w65c832: W65C832 (32 bit 6502) in an FPGA

10 points by lproven


classichasclass

Speaking as a big pro-6502 bigot, the 65816 could have been more than what it is. Two things that have always irked me about it is the use of a processor flag instead of opcode bits to determine operation width, meaning an arbitrary instruction cannot be interpreted without knowing the state of the CPU right then (is it an 8-bit store? 16?), and the segmentation model is worse than x86; you can’t even execute over the end of a 64K bank.

Unfortunately, though I expect that was for compatibility, this duplicates the same deficiencies. If this was going to be a wish list 65832, I would have hoped to see more thought made about how to get around those limitations, possibly only implementing them in a 65816/65C02 compatibility mode.