>iToddlers get BTFO 34 years laterYou can't hide your bugs forever
>motorola>emulatorok
>early cpus have undocumented instructions>now cpus are entirely undocumented and use microcode as an x86 driverwe did it
>>107497889bugs are only bugs if you know about them
Karl Jobst lost
>>107497889>tom's hardwarecringehttps://www.downtowndougbrown.com/2025/01/the-invalid-68030-instruction-that-accidentally-allowed-the-mac-classic-ii-to-successfully-boot-up/>>107501646who?
>>107501677Q: How many Pentium designers does it take to screw in a light bulb?A: 1.99904274017, but that's close enough for non-technical people.Q: What do you get when you cross a Pentium PC with a research grant?A: A mad scientist.Q: What's another name for the "Intel Inside" sticker they put onPentiums?A: The warning label.Q: What do you call a series of FDIV instructions on a Pentium?A1: Successive approximations.A2: A random number generator.Q: Complete the following word analogy: Add is to Subtract as Multiplyis to:1) Divide2) ROUND3) RANDOM4) On a Pentium, all of the aboveA: Number 4.Q: What algorithm did Intel use in the Pentium's floating point divider?A: "Life is like a box of chocolates." (Source: F. Gump of Intel)Q: Why didn't Intel call the Pentium the 586?A: Because they added 486 and 100 on the first Pentium and got585.999983605.Q: According to Intel, the Pentium conforms to the IEEE standards 754and 854 for floating point arithmetic. If you fly in aircraftdesigned using a Pentium, what is the correct pronunciation of "IEEE"?A: Aaaaaaaiiiiiiiiieeeeeeeeeeeee!Q: Did you hear about the new "morning after" pill being developed as areplacement for RU-486???A: Its called RU-Pentium. It causes the embryo to not divide correctly
>>107497889"Accuracy" of mame highlighted it? Article says the opposite. MAME's core is not accurate to real 68030 silicon and instead replicates 68020 behavior which leads to a crash. That it's a bounds fail on a jump table that triggers it is a separate issue to the headline. MAME is full of these kinds of mistakes because it's just put together by people reading the docs and pretending that's the whole story. CPUs never used to trap invalid opcodes, preferring instead to let the instruction decoder just deal with them however. Sometimes you were seeing unfinished instructions they meant to support but didn't get working, other times it's just the instruction decoding having a fit. Either way sometimes devs used them because even busted they were sometimes useful, so if you want to call yourself an accurate emulator, you need to emulate these as the silicon did and not the way the reference manual says it should.
>>107497889yeah but how does this affect me
>>107501731Thanks to AVX we now have official fast approximate reciprocals. For when you need the wrong answer really fast.
>>107497889iTODDLERS BTFO
>>107501825>"Accuracy" of mame highlighted it?yeah that made me laugh>CPUs never used to trap invalid opcodes68k series does. an invalid opcode/instruction will trigger a jump to vector #4https://68k.hax.com/ILLEGALin this weird case, the CPU sees it as a valid instruction and continues on like nothing had happened, giving the result that the apple programmer wanted but they obviously didn't see as a mistake because the code "just works" when they were programming the rom. they got lucky and they ended up at the right address to execute. they would have never seen it because the bug won't appear on 24-bit versions of the chip. if it was true 32-bit they would instantly seen the crash.
>>107502187HCF no longer existed on 68K CPUs?
.>>107502786>HCF no longer existed on 68K CPUs?no. you could safely stop the cpu or have your code execute to handle the error and return out of the current exception that the cpu is in. with systems like amiga and mac, programmers don't really need to think about it much since it's handled by the rom.
>>107501864they are wrong answers but more correct than most wrong answers which is useful
