Mac OS X Panther On A 25MHz Centris 650

Currawong writes "danamania, well known for making the most of 68k Macs, has done the ultimate, and installed Mac OS X Panther on an old Centris with 68MB RAM, a 25MHz 68040 and 4GB drive - an early 90's machine with about the same power as a NeXT cube. To achieve this, she's had to run it under PearPC on Debian, resulting in a severe performance hit, as generic emulation runs "about 500 times slower" according to the developers. On this approximately 0.05MHz G3 speed emulator, the boot screen has taken 1.5 hours to appear, and the ETA for full boot is almost exactly 1 week! Regular updates are being posted as each milestone in the boot process is reached."

Comments

Cheating?

[Hatta]

IMHO using an emulator is cheating. You're not really running it on the Centris. You're running it in a VM that is running on a Centris.

Re:Cheating?

[mvdw]

Try installing it. I've tried installing win98 on a 486/33, it barfs saying that "win98 won't install on a processor slower than 66MHz". Exact same machine, plugged in with a 66MHz processor, installed fine. Win98 also ran fine on the 33MHz processor once installed, BTW.

Bottom line is, I would guess win2k would also have these checks to make sure it won't install on a slow machine.

Wow

[bnenning]

That is impressive. And it probably even gets around Apple's BS EULA clause that claims you can only install OS X on Apple hardware.

And....

[penguinbrat]

To achieve this, she's had to run it under PearPC on Debian...

Is the excitement here that Debian ran just fine on something so old, the great work from the developers of PearPC or what it takes to get an OS to take a week to boot?

Re:Yay!

[jargoone]

Yeah, about that. I bought the RAM, and while trying to install it, I broke the fucking memory slot. :-( Now I don't know what to do. I've installed memory probably a hundred times (literally), and never broke anything. I didn't exert any more than normal pressure. I still don't know what happened.

Apple won't help -- it's explicitly excluded in their warranty. Paying for the repair would cost more than I paid for the laptop. So I'm stuck with pretty much a useless laptop, unless I go back to OS 9.

My only hope is that the logic board problem in this series will rear its head, and that they'll replace it in spite of this issue. Otherwise, I'll just have to eBay it and eat the difference.

I'm pretty bummed about the whole thing. I decided to buy my first Mac and see what the hype is sbout, and this is what happens.

Not totally.

[Inoshiro]

There are new instructions on 486+ CPUs that are not supported on the 386. Instructions like cmpxchg8, for example. Some of these can be worked around (cmpxchg8 is used for data moving, and you can "fake it" for the locking involved with more computationally expensive instructions), but some of them cannot, and either way would require extensive work in the lowest level functions of the kernel to match the differences in the design.

That's why most new packages you see are i486; they use instructions Intel added to the ISA when they released the 486.

An Excellent Demonstration of Church-Turing...

[borgheron]

This is an excellent demonstration of the Church-Turing hypothesis.

Boiled down, it basically states that any computer can emulate any other. :)

GJC

I'm not impressed

[Paladeen]

Hah!

0.05Mhz? That's just plain speedy. I'd like to see them do what I did: Run it on a 0Mhz processor: