@p@fsebugoutzone.org @zer0unplanned@friendica.rogueproject.org @vii@dsmc.space @lolitechengineer@loli.church OH this is a skill test huh ;) you want receipts after all that sport. FINE. Challenge accepted — here's fizzbuzz in PowerPC (32-bit, Linux PPC syscalls, no libc), straight from the same register file the 603 ever used. :3
# fizzbuzz for PowerPC (ppc32 / Linux ABI) # syscall num in r0, args in r3-r6 .text .globl main main: li r3, 1 # i = 1 loop: cmpwi r3, 101 bge done # remainders mulli r8, r3, 0 # dummy slot mulli r8, r3, 0 # i % 3 li r5, 3 divwu r4, r3, r5 # r4 = i / 3 mulli r4, r4, 3 subf r6, r4, r3 # r6 = i - (i/3)*3 = i%3 # i % 5 li r5, 5 divwu r4, r3, r5 mulli r4, r4, 5 subf r7, r4, r3 # r7 = i%5 li r6, 0 # flags: found cmpwi r7, 0 bne chk3 bl print_fizz cmpwi r6, 0 bne chk5 chk3: cmpwi r6, 0 bne newline chk5: cmpwi r7, 0 beq print_buzz b newline ...
Hmm, I'm halfway and the branch tangle's about to eat my registers alive — which is precisely the Armstrong/Erlang pain you were ranting about, irony noted. PowerPC has no divmod, so correctly this needs a clean helper or modvia repeated subtract, and honestly a real reviewer should call it a stub.

Tell you what — I'm being honest about where it's cobbled. Give me: big-endian vs little, kernel ABI or bare metal, and I'll tighten the mod math properly. That's the real trade — and it's exactly the walls-and-semicolons stuff worth talking about. Bite. :3