Slashdot Mirror
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."
Drill baby drill - on Mars
Why does Slashdot keep covering people who waste time installing PearPC and OSX on various already-incredibly-slow pieces of aging hardware? Is Slashdot really this hard up for quality story material?
Getting a web server to run on an Atari 800 is kind of cool. Modding a Roomba to deliver your Dr Pepper is nifty. Getting OSX to run on the slowest piece of hardware you can get Linux to run on is tired and boring.
Don't make me start reading CNN for my news.
That's the stupidest thing I've read all day long. And I've been reading POLITICS all day long.
Of course, people always ask "Does this have any practical use?"
Absolutely not. But that it not the point. In The Real World imagination and creativity are the driving force. How do you foster that? By challenging yourself and inspiring others. There does not have to be any realistic application as much as there needs to be a thought process behind it that can be capitalized on in the future. Experiments such as this drive the imagination and the mind into new directions and those new paths we explore can lead to really, really utterly brilliant things that can have a profound effect on our lives.
In school, a teacher once told me "Answers don't really matter at all. The process you use to reach your conclusions is the most important thing in the world." It blew my little mind open to the true nature of creativity and for the first time I valued it in a way that was truly profound.
I wouldnt think a Linsux user would be inviting comparisons of boot times... Win XP is POST to desktop in 14 seconds on my machine. Linux takes 50 seconds.
Or just make a new category: "Useless". This is when you've run through the main page and the several sections that are somewhat important. I spent a week and a half installing gentoo on an old computer, but I don't think it should be posted on slashdot.
If you're using hardware, I'd argue that it is native. Anyway, I find it far more impressive that debian runs on this machine, than OS X "runs" on this machine.
Give me Classic Slashdot or give me death!
You're missing the point - Mac OS X doesn't run on the old 680x0 Motorola processors natively AT ALL, only on the newer Motorola PowerPC processors (Apple only supports G3 and higher, but I seem to recall the original developers' platform used the 603/604 chips).
:)
PearPC, however, will compile on a variety of Linux platforms, which will emulate a complete G3 Mac. But, obviously, the emulation is going to be ridiculously slow on a chip several generations behind the current
So, this is more a case of "because I can" rather than anything practical.
Flamebait? C'mon, people, get a clue!
Using an emulator does count as cheating.
If I run Bochs to boot Win2k3 on an old 386, will I get a Slashdot FP?
Shit, I suppose I can look forward to seeing that tomorrow now, can't I?
I appreciate emulation, I really do. But aside from the author of the emulator, no one else gets to claim geek cred from using one. Had this person really gotten OS X to run on a 68040, I'd consider it somewhat cool. Running it on an emulator, though? Here, hold on, I'll come back and describe my experience getting SMB3 to run under SNES9x on a 2GHz Win2k box... Woo-woo, rolling in the coolness now, baby!
I feel sorry for people who bitch about how this has no "practical" use. I can't help thinking they're the same ones who walk into art museums and make winning comments like "pfff, I could do THAT..."
vk.
What exactly would stop Win2k3 from running on a 386? The ISA is the same...
My other car is first.
What's wrong with Marathon and Spectre? That's why I'm keeping around a couple of my beige Macs - a Q840AV and G3/300 to be exact. There a dozens of fan-made Marathon scenarios I have yet to play. In terms of storyline and gameplay I still think the Marathon series was the best FPS I've ever seen.
Constitutionally Correct
A more accurate comparison would be to run XP on a 486/25 with 64 MB of ram. Of course, XP will probably refuse to run on a 486 at all, so you'll need a 686 emulator running on the 486, and you'll need at least 128 MB of ram (so the emulator will have to use virtual memory to emulate the extra 64 MB + that used in overhead.) I have no reason to expect that if the emulator is good that this won't work.
It'll probably run faster than MacOS X on the Centris too. After all, OS X needs a PPC, which is totally different than a 680x0, so it needs to be emulated at the lowest level. But a 686 isn't very different from a 486, so an emulator could take advantage of this.
That this works at all is not really a testament to the robustness of OS X, but instead a testament to the robustness of the PearPC emulator. As far as OS X is concerned, it's running on a PPC box. Just a very slow one ...
Actually I beleive the difference between Centris and Quadra was more marketing than technology. They were essentially the same machine. Boot times on the 68xxx apples could be improved by putting system files on a ram disk, which was a virtual disk maintained by AppleRom. Great featuer
I'm not hot for her because she's a chick. I'm hot for her because she's a chick who gets modded up.
Oh, really? Than you should think about this:
Lameness filter encountered. Post aborted! Reason: Please use fewer 'junk' characters. Hopefully my explanation will dilute those "junk characters" and will let me post this comment. It's interesting that this lame filter stops me from quoting programs but doesn't stop anyone from posting full-screen ASCII-art swastikas and pornography. But anyway...
Thanks to the Inline module, it is possible to include fragments of C code in Perl programs. You can write part of your Perl program in C (for example one speed-critical subroutine) and it is automatically compiled to native binary machine code and linked as a shared object (see this comment of mine and read the paragraph starting from "Actually, inlining other languages..."). CPR stands for "C Perl Run." From the description:
In other words, CPR program is a C program which is run by Perl, just as if it was a C code inlined in a Perl program.
Now, in this case, the C program I quoted (which is itself run by Perl), includes a Perl code inlined in C by CPR_eval(). What is inside that inlined Perl code is an inlined C code (use Inline...) which is a C function greet() that returns a C pointer to C string "Hello world". The next part of the original (outermost) C program is a C printf() function printing two C strings. Those C strings, arguments to printf(), are returned by two invocations of CPR_eval(), both of which inline Perl code. The second one just returns Perl interpreter version, but the first one is more interesting. The first CPR_eval() returns a C string to printf() which is converted from a Perl string returned by the Perl code inlined in that CPR_eval(), which is a call to Perl greet() subroutine which was defined earlier by the C function inlined in the Perl code inlined in the C code by the first CPR_eval() invocation. It all happen inside a C main() fu
Sincerely,
Pan Tarhei Hosé, PhD.
"Homo sum et cogito ergo odi profanum vulgus et libido."