Well, I can say it truly is a miracle. I never thought I would be able to install Mac OS X 10.3 on my supposedly “inferior” x86-based AMD machine. Well, for most of you mac-zealots, you were wrong. It actually does. And I’ve got the proof right here!
The installation was not really that hard; the PearPC website has done an excellent job in providing you with information. And, just days after the 0.1 release, websites all around the world posted howto’s and installation guides. I’m not going to explain how it is done in this article; that is not the purpose of this article. The purpose of this article is to simply show you that it can be done, and I will do this through a number of screenshots.
After generating .iso’s from the Mac OS X install disks, it is time to boot PearPC by issuing a simple “ppc config.txt” command. PPC itself does not really have a GUI (only the “change CD” button) so the first GUI stuff you’ll encounter is the Installer:
My hart actually skipped a beat; I never really thought it would work; not necessarily because I thought PearPC was vaporware, but more because my computer is getting kind of, well, old.
Of course everything was not running very snappy; on their website they warn you: the emulated processor is about 40 times slower than the host processor. Still, I was amazed at what I saw: it worked!
After a painstaking process of creating a bootable “harddisk” (grab a bootable .img file from the website; it saves you a lot of time), the installation started.
It took, in total, more than 5 hours. And that was just the first cd. Since I had nothing else to do (PearPC took 99% of my processor and all the RAM it could possibly find), I actually started to clean my bed/computer room. Thank you, PearPC.
After the installation of the first disk, PearPC had to be restarted with the harddisk image file as the master, and the second install disk as slave. This was a point of frustration; Mac OS X would not boot. Luckily, I was not the only one having this problem. It turned out that the installer put the Mac OS X partition in third, while it should be in second place. Again, take a look at the PearPC website, the solution is out there.
After the installation of the second disk was ready, I was finally able to boot into Mac OS X 10.3, Panther. This, ladies and gentlemen, is a very strange moment. I now know what all those people must have felt when they were abducted by aliens (not that I believe in those stories, but still).
Then, the big welcome:
Spotting the memory/processor usage:
Creating a user account:
And, finally: Enjoy your Apple computer!
Well, just when I was ready to start “Enjoying my Apple Computer”, a weird bug appeared: I was dropped into an infinite Finder crash-loop. Well, not really infinite, 15 minutes or so later, the loop ended. Mac OS X 10.3, Panther, was actually running on my Athlon machine!
The only glitch left was the Dock. It also encountered a crash loop, but this one was indeed infinite. I am confident that these problems will be resolved in later versions of PearPC.
Conclusion:
I am truly impressed. I do not really have anything more to say than: thank you, PearPC, for this, well, enlightened moment. And for my clean room, of course.
Test system:
– AMD Athlon XP 1600+;
– 512 MB SDRAM;
– Ati Radeon 9000 with 128 MB DDR-RAM;
– CMI-8738 based 5.1 soundcard;
– MSI K7T Turbo2 mainboard;
– 40 GB harddisk;
– Standard ps/2 keyboard;
– Microsoft Trackball Optical (USB);
– Windows Server 2003, Enterprise Edition (host);
– PearPC 0.1, emulating Mac OS X 10.3 Panther (client).
About the Author:
Thom Holwerda is a regular visitor on OSNews.com and has contributed more than once. His first computer experience dates back to 1991 (a 286 entered the household). Over the years he has played around with several computers, but it wasn’t until 2001 that he really started to experiment OS-wise with computers. His favorite operating systems are Windows Server 2003, Mandrake Linux, BeOS. He also has an affinity for the QNX Neutrino RTOS. Thom also contributes to the SkyOS project, being responsible for the Dutch translation.
Technically interesting, but having an OS running far slower than native, and having to pay both the Windows and Mac OS X licenses to run it does not sound like any useful nor economically rationale.
If you want Windows, buy a PC, if you want OS X, buy a Mac; emulation is just a technical curiosity with little usefulness in most cases, other than raising your geek factor when showing to your peers (chicks will not be impressed, I’m afraid)
Well, the main reason to emulate anything is to see/prove that it can be done.
Emulating a PPC on x86 hardware is impossible, as everyone knows 🙂
Now they prove the world wrong, and what do people say? “It’s soooo slow, it’s not worth the effort, you’re better off buying a cheap G3 mac”.
Bleh. These projects are HOBBIES. People have hobbies for FUN.
The fact that they like to share the results of their hobby with the world is only positive, so stop with the negative attidude already.
I’ve already mentioned some of that stuff when PearPC came up for the first time, but it seems a few people haven’t read those.
QEMU
Those hoping that QEMU might improve performance, forget it. If I’m not hugely mistaken then QEMU is designed to be easily retargettable, and part of that design is that covers (ie. the native code that is used to translate instructions) are actually generated by the C compiler, which is a neat idea.
The problem is that such compiler generated code will never be as efficient as hand-coded assembly when it comes to translating from one machine language to another.
JIT, dynarec, binary translator, or whatever you might call it[/i]
The 1/40 speed is with the included JIT enabled! According to the PearPC website the speed with the interpreter will be about 1/500:
http://pearpc.sourceforge.net/about.html
There is still some space for optimization, but even the authors of PearPC said in a German interview that they hope to achive 1/10 of the speed, which sound reasonable, but I doubt that it’s going to be faster than that:
http://www.macnews.de/index.php?_mcnpage=52752
AltiVec
Even though OSX runs without AltiVec, emulation of that instructon set might speed up the OS and several applications.
But with AltiVec the difference between PPC and x86 is even more apparent than with SISD. An AltiVec instruction can use 4 of 32 registers, while SSE can only use 2 of 8 registers, or in the case of x86-64 2 of 16 registers. That’ll be very tricky to get that running fast…
PearPC running OS9?
” So far no one has succeeded in doing that.”
http://pearpc.sourceforge.net/faq.html
I don’t see PearPC as a threat to Apple, because it’ll actually make Macs look better than worse, since even the fastest available PC will be able to emulate a half-decent Mac at full speed at that time.
it is irellevant that it is slow… the whole deal is that it can emulat POWER PC HARDWARE in a working form.. its RISC based.. a completely diffrent instructionset to slow X86 machines.. and that is a first for anything available to normal people..
Now if only BochsX86 and PearPC people could join forces we could perhaps one day have an actual useable program instead of typical OpenSource abandonware..
This technology could be useful in a WINE type project, but allowing Windows/Linux to run native MacOS X applications.
This technology could be useful in a WINE type project, but allowing Windows/Linux to run native MacOS X applications.
If someone can do this really good, we can say good bye to Apple. (I don’t think this’s anything good, dude.)
“If someone can do this really good, we can say good bye to Apple. (I don’t think this’s anything good, dude.)”
well this ain’t gonna happen. soone or later apple will comup with something cool that everyone will be playing catchup with. remember all the companies are still trying to make their mp3 player as cool as iPod & same is the case with iTunes
To my understanding PearPC emulates PPC architecture, WINE is an implementation of some of the Windows API on Linux. One is emulating hardware, the other isn’t really emulating, but is moreso implementing (and not even hardware, software).
WINE == Wine Is Not an Emulator
and it’s not… don’t ever forget it!
> x86 CPU’s are never going to be able to perform
> exceptionally at PowerPC emulation for one VERY SIMPLE
> reason… registers.
That’s simply not true. It’s like saying that any program will never be faster on x86 because the ppc has got more registers: clearly something untrue. When you run a JIT compiler, the PPC code becomes just yet another language to compile in the native CPU’s language.
Most of the times, PPC code uses only a minimum fraction of the registers the PPC has, most of the other registers are used only as an extension of memory, because the only operations the PPC can do on memory are load and store operations. Obviously, since the x86 doesn’t have this restriction, it can avoid having more registers and use the memory instead.
For example, x86 processors are known to be pretty good at emulating 68k processors, yet the 68k ones have double as many registers as the x86 ones.
> Yeah, except the OSX license agreement explicitly states
> it can only be run on Apple labeled computers. Bong, try
> again – this is a violation of the license.
Which, in itself, is not illegal. Also, such clauses are invalid in many countries.
it is irellevant that it is slow… the whole deal is that it can emulat POWER PC HARDWARE in a working form.. its RISC based.. a completely diffrent instructionset to slow X86 machines.. and that is a first for anything available to normal people..
First of all, RISC emulation has been done before: the MIPS R3000A and R4300i CPUs in Playstation an N64 emulators.
Secondly, PPC emulation also has been done before, by a modified SheepShaver.
Thirdly, are you suggesting that Mac users aren’t”nomal people”?
Anyway, the only “first” thing of PearPC is that it emulates a PowerMac more or less fully, while SheepShaver is more like a runtime system with a PPC emulator stuck onto it.
This technology could be useful in a WINE type project, but allowing Windows/Linux to run native MacOS X applications.
That’s just rubbish. WINE doesn’t emulate *anything*! It’s a combination of libraries that mimick the Win32 API, and a PE loader to be able to run Windows executables.
No hardware emulation is involved, because the application code runs directly on the CPU (which is why Darwine won’t run on MacOS X until they’ve added a x86 emulator), and Windows applications cannot access hardware directly, so everything has to be done through the API anyway.
Please learn about what these projects actually do before you suggest that they would be great in combination with something else.
On registers being a limiting factor
That’s simply not true. It’s like saying that any program will never be faster on x86 because the ppc has got more registers: clearly something untrue. When you run a JIT compiler, the PPC code becomes just yet another language to compile in the native CPU’s language.
You fail to see the difference that there is a difference between a compiler generating optimized code for a high-level language and directly translating machine code to another machine code.
Sure, you could see the load/store architecture as having a disadvantage, because it only operates on registers and has to use special instructions to access memory. But just like x86 code is optimized to hide its problems, so is PPC code, ie. most of the code will try to work with lots of registers, and each instruction can use 3 registers at a time, while x86 can only use 2 registers or 1 register and 1 memory location per instruction. Since PearPC is not generating code for some high-level algorithm but to reproduce what the PPC instructions do, you’ll find the x86 have to perform a lot more memory accesses.
I’m pretty sure that the emulated register file fits into one of the caches, but accessing the cache is still slower than having the value in a register.
Of course emulation of RISC processors with 32 registers has been done on the x86 before, but in that case it was the relatively simply MIPS arcitecture, and the R3000A in the Playstation is only clocked at 33MHz and the R4300i in the N64 is clocked at 96.75MHz (IIRC). Also the R3000 has no FPU, and although the R4300 has one it’s still not quite as superscalar as a G3 or is clocked with the speed you’d need to run OSX fluently. Not to mention that PPC is way more complex than MIPS is.
For example, x86 processors are known to be pretty good at emulating 68k processors, yet the 68k ones have double as many registers as the x86 ones.
But like x86 the 68K only has two-address code, and comparing a PowerPC to a 68K is a bit far fetched: the fastest (and most primitive) 68K instruction takes 4 cycles to execute while a PPC issues 3 or 4 instructions of that complexity per cycle and it’s also only clocked 20 times faster if it should be anywhere near usability for OSX.
That’s just rubbish. WINE doesn’t emulate *anything*! It’s a combination of libraries that mimick the Win32 API, and a PE loader to be able to run Windows executables.
What do you think an emulator is? Or do you think GNU’s Not Unix as well? (well, I guess since there isn’t a full GNU kernel yet…). If you emulate (read: mimic) the Win32 API and a PE loader, you’re more or less emulating Win32.
No hardware emulation is involved, because the application code runs directly on the CPU (which is why Darwine won’t run on MacOS X until they’ve added a x86 emulator), and Windows applications cannot access hardware directly, so everything has to be done through the API anyway.
Of course, Win32 itself was created to abstract the hardware, so x86 emulation should be completely irrelevant when it comes to making Windows software run on any platform. The only reason Mac users use x86 emulation to run Windows software is because it
a) allows them to run the complete environment with the actual Win32 API, explorer, etc for better compatability; and
b) allows software that does rely on the x86 hardware to run properly
Remember, Windows NT (upon which 2000 and XP are based) was originally built for MIPS with hardware abstraction as a specific objective (which allowed them to run on x86, PowerPC, DEC Alpha, etc). If you emulate the API properly, most applications will run fine. If you emulate the x86 processor well enough, though, you can install Windows without emulating it and not worry about it.
Please learn about what these projects actually do before you suggest that they would be great in combination with something else.
He simply mentioned that it would be great in that type of project, not that they be combined. The only real problem is that they offer different goals. On the other hand, if running OS X applications on Windows/Linux turns out to be problematic because they expect the hardware to be PPC, then hardware emulation could be a very important part of any project that attempts to do the same thing WINE does for Windows applications.
I understand how WINE works. I am not a programmer but rather than just intercepting MacOS API calls from a loaded binary such as WINE does, it would translate the binary into X86 using the code from PearPC, load it, and then intercept the MacOS API calls to bring up the MacOS program on the x86 OS.
Most of you don’t get it…at all.
PearPC is not about usability.
PearPC is not about being a cheap alternative to owning a mac.
PearPC is about them getting_to_run MacOS on a damn x86.
PainKilleR:
What do you think an emulator is?
The term emulation was invented in the mid-60s by IBM when they realised a 7090 simulator in S/360 microcode. Since it isn’t possible to modify the micocode that easily these days and in terms of the law microcode is the same as software, for me an emulator is a piece of software that simulates another computer system, including the CPU.
That’s why WINE and VMware are no emulators. The latter is a virtual machine, and I don’t think there is a simple term to describe what WINE actually is.
Or do you think GNU’s Not Unix as well?
But GNU isn’t Unix. Unix is a trademark for a special operating system that was originally designed at the AT&T Bell Labs, while GNU is a bunch of software of which some might be similar to tools Unix has.
But when even the creators of WINE claim that “WINE Is No Emulator”, why do you think that I’m wrong saysing so too?
If you emulate (read: mimic) the Win32 API and a PE loader, you’re more or less emulating Win32.
You might it call a Win32 simulator, but not an emulator, because emulation certainly involves CPU simulation.
Of course, Win32 itself was created to abstract the hardware, so x86 emulation should be completely irrelevant when it comes to making Windows software run on any platform.
Erm, you do know that programs contain not only calls to the API, but actually machine code to do calculations or whatever the application is supposed to do?
And BTW, with all that abstraction FX!32 still needed so-called “jackets” to transform x86 API calls to Alpha API calls, because parameter passing can be quite different on various processors as well. And in that case the API didn’t even have to be ported, because the Alpha already ran WinNT.
Remember, Windows NT (upon which 2000 and XP are based) was originally built for MIPS with hardware abstraction as a specific objective (which allowed them to run on x86, PowerPC, DEC Alpha, etc).
Yes, but that only meant that WinNT was more easily portable to these platforms, but applications still had to be ported to the platform as well, even if the porting process is just a compilation of the source code. If you don’t have the source code you won’t be able to run a WinNT/x86 application on WinNT/PPC, same API or not.
That’s why there was FX!32 for WinNT/Alpha: To interpret the x86 CPU and then do a partial static translation of the WinNT/x86 binary.
And that’s also why FX!32 is an emulator, because it emulates another processor in software, despite the fact that both systems have the same API.
Dave, you wrote:
This technology could be useful in a WINE type project
Since WINE includes no emulation, but running Mac binaries on a PC requires at least CPU emulation, no matter if you have the API ported or not, this simply cannot be a “WINE type project”.
Of course it would be possible to combine a CPU emulation with a port of the API, which is quite similar to what Darwine tries to achive in the future the other way round, but that doesn’t really have that much to do with a “WINE type” approach anymore.
WINE is a port of the Win32 API to Linux and a PE loader, because Linux can only handle ELF binaries itself.
What you are suggesting is an emulator that is combined with high-level emulation in form of the API.
Since WINE includes no emulation, but running Mac binaries on a PC requires at least CPU emulation, no matter if you have the API ported or not, this simply cannot be a “WINE type project”.
All I meant by “WINE type” was that it runs binaries intended for a different platform than the host… as I am sure you are aware.
I am a programmer. I write for windows mostly and for the Mac whenever I can. I prefer the Mac, and would gladly pay the $100-200 premium for a Mac. However I love seeing this, and will install it on my windows machine, just to see it work. But it is a pure “for fun” kind of thing. 40-500 times slower certianly translates to useless, but it is fun.
I suspect that eventually they will achieve something like 10 times slower. As PC’s start to approach the 4GHz mark, this will be a emulator that will run like a minimum OS X configuration and perhaps get some windows people intrested enough to make a switch.
However, they may discover as I have that a Mac OS X machine running Virtual PC can do more. My VPC benchmarks as a 500MHz PC, while running on a 1GHz Mac – that is pretty incredible. I can’t play games in VPC, but most other programs run slowly but are usable.
If you try this and like OS X – it is just too slow, start saving for an eMac. You will find it to be a wonderful alternative, and you can do your windows work relativly well (So far I have only found one program that did not run on VPC – but i didn ot even bother to try games).
PC’s are great game consoles! Better than a PS2 and only slightly more expensive.
Look its simple the idea behind pearpc is to make a PPC cpu emulator, due to its well made code it is fortunate enough that it does emulate a G3 computer and permits you to run macosx at for now ultra low speed (still you have to keep in mind that is in an early stage “v0.1.1 or v0.2pre”)but it does also emulate other PPC os, and dont worry about Apple going bankrupt there will always be a huge amount of rich people that will stick to there oh so powerfull G? to be the proud owner of one of those expensive jelly bean machine.
I mean look if i would be able to make you a ford tempo that as the look of an audi and you would want an audi would you buy my look alike or the true stuff.
so i really dont think that its a big problem for mac or it would be the same for windows with all those mac capable of emulating X86.
“Hmmm, is there actually someone honest on Ebay selling Macs? I’ve heard way too many horror stories about that site, seen too many people get burned…”
My uncle/aunt, cousins and niches (?) buy regulary on eBay. You can find yourself some way cool stuff over there for cheap prices but you also have to be prepared for assholes: people who sell (semi-)broken stuff while not mentioning the actual state; some do state the hardware is partly broken which according to my observations reduces the price. Those who don’t have the best change when they state that there’s no guarantee and then claim some story the hardware got broken during the sending. It doesn’t mean everyone who doesn’t give warranty is doing this, but it is a risk.
So how do you circumvent that? First of all, buy from one who gives warranty. That’s mostly 8 days (a week). According to EU law one has to give 1 year warranty and if that’s not included then one has to state that. So when not stated you are able to fall back on that. Second, another way would be to get the package yourself from the person who’s selling. Test it there. Happy? Buy it. Not happy? Don’t buy it, and damn the transport costs you’ve made. Third, a way to find out wether a buyer/seller can be trusted or not is to look at his/her history, see what % of people were happy, and see what the resons of the people who were not happy are.
According to my observations the first option raises the price. People tend to pay more for a product if it includes warranty. I’m currently trying to buy a SUN Ultra 5 and shit happens when people pay as much as 120-200 EUR for it, when it includes a few days warranty.
However it has to be said they buy from eBay.de. They live in Germany (on Dutch border) and eBay.nl is still small. I only have made one deal via eBay (a trade, not a sell) and i’m moderately happy with it since a lil bit of the hardware was broken which wasn’t stated. Then again, it also included some hardware which wasn’t mentioned.
I’d say: be careful on eBay, but don’t dismiss eBay or the concept of eBay right away. Trying to get a relation with the seller also creates trust and might give you a better image of the person. Counts for the seller creating rel. with buyer as well.
Happy hunting =]
As for PearPC. It is merely a working concept. Slow as hell. Whoever buys MacOSX to seriously use it with PearPC is quite nuts, imo. Because i don’t want to pirate, i leave this alone. There’s a technical reason why emulating PPC is downright _practically_ impossible on x86, but i don’t know it.
I don’t know if you saw, but PearPC 0.1.1 was released and it fixed the Finder crash-loop
I’m booting OSX fine on my p3 500mhz. It isn’t as slow as people claim it is.
if you read the license agreement for OS X you shall see that it is illegal to run OS X on a non Apple-labeled computer. It is also illegal to run it on more than one Apple-labeled computer.
And by “Apple-labeled” I would assume the text that is written on the machines (underside of machine on notebook Apples, not sure on desktop systems)
Anyways, PPC CPU’s have on-chip Endianness conversion, whereas intel chips have to do the conversion as code. So until x86 based chips get on-chip conversion to do Little Endian to Big Endian (and vice versa), then PearPC will ALWAYS be slower than VirtualPC on a Mac
Allowing to install that one-mouse-button P-of-S OS on perfectly good x86 hardware is a crime against common sense.
MacOS on x86?! Why spread a sickening, brain-paralyzing affliction on otherwise normal PC users? What a waste of programming skills!
Look what Mac did to Slashdot – it became Macdot of sort. I am not joking. The world is going to end.
I’ve been using x86 stuff since the 286 and my only experience with macs have been brief moments of using them at school. I’ve never used Mac OSX, but I’ve studied all of the videos on apple.com to get an idea.
Everyone is always trying to sell the notion that if you switch from a PC to a Mac with OSX that you suddenly become enlightened to a whole new world that you never knew existed. Well that’s a nice idea, but it just so happens to be Apple’s marketing message… so my belief is that most people are just too susceptible to marketing tactics.
One day I decided I wanted to try out Mac OSX to see what the fuss was about. I did some research and found out that there were no emulators that could emulate OSX, so I downloaded one for OS9 and tried that. It was obviously old and outdated, so it wasn’t going to show me much. The next step was to ask some Mac people if they’d let me VNC into their machine to test it out… well the Mac people made fun of me and even after several attempts, I was unable to find a willing participant.
“Just go to the store and try one out”. Ya, but I don’t go out… I spend most of my day in front of the PC and I don’t like to go out, especially if it’d be only to test a computer that I used to hate using when forced to in school.
I thought that if apple.com could have some servers setup, running OSX on them, and allowing a remote desktop type interface from their website, then this would greatly open up their exposure to people like me who are somewhat interested but not interested enough to go out of our way to experience OSX.
Finally I gave up, and when it came time to buy another PC I briefly looked at the G5s, but since I had no OSX experience I quickly ruled them out. Instead for a little over a grand I built an Athlon 64 3200+ with 2GB of ram using the finest components available.
Now I see this, “Is it a miracle?”. Yes, it definitely is! I’m very curious and I can’t wait to try Mac OSX. I don’t care if it’s slow or buggy, the whole point is just to familiarize myself with the platform.
Apple should definitely allow these projects to continue. They are not a threat to Apple, but instead allow more exposure to their flagship OS.
I’m going to go give it a try. My hypothesis is that despite the articles, “I was a lifetime intel user, but then tried OSX and found God”, I am not going to be impressed enough to justify puchasing a Mac. I hope that I am though. Most of my software runs on the Mac platform, and I’d love to have a better OS to play with… but I’ve been using Microsoft products since very early DOS days, and I’m so accustomed to them that I don’t even have to think anymore to use/fix their software. Plus, I’ve put up with so much crap when Microsoft software sucked, and am only recently able to enjoy solid/stable software, that I feel like I need to recuperate my losses.
According to C’t (“Nachschlag zum Zuschlag”, 19/2003) that disclaimer (when bidding you agree to not have the one-year warranty as prescribed by EU-law) is complete nonsense. There is no such EU law, and therefore the disclaimer is nonsense. May I quote:
“Eine zwangsweise zu leistende zweijährige Funktionsgarantie gibt es nicht, auch nicht nach EU-Recht. Und um die Verantwortung für das, was er liefert, kann sich der Anbieter auch durch Disclaimer nicht herumdrücken. Er bleibt verpflichtet, seine Ware zutreffend und sorgfältig zu beschreiben, und darf auch nicht absichtlich Irrtümer bei Bietern provozieren.”
In other words: the seller must describe the “actual state” of the product clearly, completely and carefully. The EU Law disclaimer does not change that.
I also know someone who has bought some things on the German eBay. Most of them were perfectly working and for the only cheap non-working PC he eventually got his money back.
Besides that, I don’t know how popular Ebay is in the UK, but in the Netherlands you might also want to take a look at http://www.marktplaats.nl/ or in case you are looking for a Macintosh http://www.tweedehandsmac.nl/
I’m presuming this is possible, but I’ll ask anyway …
We’ve got some new Linux servers at work, they’re quad-processor 2.4 GHz Xeon boxes that run Red Hat Enterprise Linux WS 3.0 Update 2. Seems like a good platform to maximize PearPC 0.1.1 “performance” on, oui? 😉
But, the catch is, they’re headless. Can I just get on them and fire up PearPC in/as a remote X Window and do the install that way, and run it remotely? (Yes, obviously there’ll be a hit compared to running it directly on a box with a head on it. But we’re using switched 100 Mbit everywhere – in some cases 1 Gbit – so the network overhead isn’t too bad.)
Allowing to install that one-mouse-button P-of-S OS
This alone shows that you are just a plain troll who doesn’t know jack about MacOS X, because it supports three buttons and the mouse wheel right out of the box with no additional driver installations. And unlike the IBM Thinkpad with WinXP my iBook with OSX doesn’t forget that it has a mouse wheel and a third button every time I wake it from sleep.
on perfectly good x86 hardware is a crime against common sense.
One must really have lost all objectivity to call x86 “perfectly good”, which not only according to my own impression is one of the worst architectures ever.
Unless PearPC uses several threads you won’t see any real speed-up compared to a single-processor machine, apart from the fact that you still have reserves for other applications.
And even if PearPC does use threads, the CPU emulation won’t be able to take advantage, because it simply has to run in a single thread. There is no way I can think of the split such a task into several threads.
I know one of he speed up tricks that BasiliskII (a 64k mac emulator) used to improve emulation speed was to intercept some of the more common operating system functions like calls to certain libraries. Instead of executing those functions in emulation it ran x86 native versions and then fed the returned value to the emulator. Of cource this limits BasiliskII to only running MacOS, but I wonder if it would be possible to utilize the work done in GNUStep with PearPC to create a faster Mac specific emulation.
It seems to me that the issue that PearPC helps solve is to give a x86 users a method for trying out OS X for a lengthy period of time, and using the convenience of their own PC.
If you are a Windows user, and want to try a Linux desktop, you can easily download something like Knoppix, burn a CD, reboot, and off you go. Reboot and you are back in Windows.
If Apple would provide something similar for doing this, then it seems like this giant gap in trying to attract new Mac users would be gone.
I know personally, I am seriously considering a switch to a Mac, but I don’t feel comfortable plunking down money for a system that I can’t demo in a reasonable manner over a reasonable period of time. I’ve stood countless times in Apple retail stores playing with OS X, but the experience isn’t the same as if I were sitting at my desk at home.
The overall cost of a Mac isn’t something that keeps me from switching, it’s the “unknown” factor. If I had an easy way to demo the Mac OS X experience, I’d feel much more comfortable buying one.
I thought the acticle was about people who are trying out there programming chops by making a PPC emulator capable of running OS X.
I don’t think it was intended to say, “Na na na, you Mac Goons have got it coming to you now that we have the Panther too!”
It’s just people saying, “Well, check this out. This is what we did the other day…”
Good on them.
ps. I’m not giving up my TiBook.. 🙂
“According to C’t (“Nachschlag zum Zuschlag”, 19/2003) that disclaimer (when bidding you agree to not have the one-year warranty as prescribed by EU-law) is complete nonsense. There is no such EU law, and therefore the disclaimer is nonsense. May I quote: […]”
Daan, thanks for pointing this out!
“Besides that, I don’t know how popular Ebay is in the UK, but in the Netherlands you might also want to take a look at http://www.marktplaats.nl“
I wouldn’t advise that site. It ain’t an auction site (can be an advantage), it is easy to circumvent the site, and IMO it is rather expensive (regarding old-high-end hardware — imo). Another popular auction site in the Netherlands in http://www.ricardo.nl (afaik the most popular one). There’s a few others like http://www.qoop.nl never used either of these.
The site markplaats.nl was in the TV program Kassa last week or the week before that, where some issues were raised. The spokesperson argumented quite well, but i can’t remember the discussion anymore.
This is a good accomplishment worthy of praise. It may not be fast or useful at present, but PearPC is at the minimum an excellent “proof of concept”. Thanks for the brief article (I had not heard of this project).
I have heard several people claim we’d never see OS-X on x86 hardware. Emulated or not, the underlying hardware is “x86”
This could lead to an intersting future. I’m sure Apple’s legal team is already planning for the day they will step in and attempt to quash this. If the genie gets too far out of the bottle, and one day OS-X can run efficiently on non-Apple hardware (not necessarily x86, but in general), it could create problems for Apple’s already tight hardware market.
mac zealots – they know who they are. If you’re just a mac user and are offended by the term, grow up. I use x86 hardware, different OSes, and don’t feel offended when the term Windows or PC zealot is thrown around (of which I am not, and am quite comfortable with myself).
I think this, and any other such project, is wonderful!
Without trial-and-error, off-the beaten-path, “I wonder if” projects like this by intelligent, curious, open minds we’d still be trying to kill dinner with a rock!
If you don’t like it, don’t try or use it. If you don’t like x or y computer (or anything else for that matter), don’t user it. Sure, I may gripe about Microsoft, but it’s not the company or it’s many employees that I dislike. Many are exceptional people. It’s about the overflow of EGO that’s been allowed to dominate company policy there for so long. That’s an ethics matter that only concerns a few (or less) individuals, not the rest of the people, software, or any machine it runs on. I think it would be a good experience for many of the negative attitudes i’ve read here to actually USE the machines and software they’re bitching about for a while so they can learn that each has it’s good and bad points.
“…
And even if PearPC does use threads, the CPU emulation won’t be able to take advantage, because it simply has to run in a single thread. There is no way I can think of the split such a task into several threads.”
One processor can run the recompiled code while the other can recompile and optimize. There will be some problems with latency when encountering new code paths so the first (stupid) recompilation should probably be implemented on the “runner” cpu and let the other one do more advanced compilation later.
To get good performance out of the emulator on the x86 one would need some pretty advanced optimizations so such a divided approach can make sense. One problem would be how to let the optimizing processor get profiling information from the “runner” processor without delaying too much…
I’m just thinking modularity. Would it not be possible to have multiple instances of virtual processor threads to run on each processor, and then make the guest OS aware of these virtual processors and to do with them what it likes?
(I’m no programmer (well, not really) and certainly don’t have any experience in emulation, but I do like to think about such things and speculate.)
“While I agree with that, PearPC runs OS X so slow because it doesn’t do altivec.”
My G3’s don’t do altivec either.
I see alot of negativety from people…does anyone actually realize how good this project is…I remember when people used to say this imposible that imposible and i knew that there nothing imposible and so did alot of people…I wish i had the skill these developers have to prove alot of things wrong but programming is not my thing…I mean how impresive is this…so called experts said this is imposible…what bullshit…i installed macosx on my athlon/linux box…slow yes…but surprisingly bareable…when this thing gets optimized it will become soo much better…im looking at the initial release and im like wow…and those who claim its imposible to make it useable are wrong just like the ones who say imposible…I bet this will blow alot of things out of the water and prove all the closed minded narrow thinkers
in it’s place…it’s very good stuff!
I have a relatively modest, but current PC w/a 2.4 ghz celeron and 512megs of ram. I got Mac Os x 10.3 installed in about 3.5 hours. A friend of mine (who just switched to linux btw) came over with her copy of MacOS and we gave it a shot just for fun. We followed the Darwin-less install procedure from emaculation which worked w/o a problem. Really it was much better then I expected from all the ranting in here about speed. In fact, if you have ever used bochs its quite comparable. PearPC running MacOS was actually a bit better then bochs running knoppix. Color me impressed.
I would love be able to run Mac OS X on my PC. I think that X86 CPU’s are much faster than PowerPC G4. If they can get the emulator to translate instructions to SSE2 it would be right up to snuff with G4’s maybe better. My G4 500Mhz crawls comapred to my Athlon system.
Forget emulation… Apple should port the beast.
I’ve seen time and time again, Mac users that tout the stability of OS X because of the software and hardware ties.. You need to stop that stuff yesteryear. First, XP is very stable and so is W2K with lots of hardware. Linux runs on multiple platforms and is very capable of running most hardware without manufactures drivers. Linux is the king of OS stability. So the old argument back in the days of Pre OS X is long gone with the hardware/software marriage.
If RedHat and Susse can make it as sales and service software companies, I don’t see why Apple can’t.. A corporate buyer will not want to tie themselves into one hardware software vendor across the board. Many individuals want choice too. Enthusiasts want to do upgrades.. Ever try upgrading a Mac? I have.. and it’s not cheap and you run the risk of instability… I took my G3 400 to a G3 800.. .it was expensive.
I imagine that a lot of PC users would be happy to give OS X a try for $129 + $49 for iLife as apposed to throwing down a lot more for a new or used machine. Remember Apples market share is shrinking and Linux on the desktop is getting better. It’s just a matter of time.. Apple should make its move now.. Especially before LongHornyShit is available.
Oh yes.. and I’m a Mac user.. Just ordered my 15” PowerBook 1.5Ghz and Radeon 9700 128.. But I see through the reality distortion field of Apple headquarters.
Irving, don’t count on it. Sure, using some of x86s extra features would speed up the emulation. a bit, but it isn’t going to make the emulator an exceptable desktop envirnment. If the guest envirnment is 40 or even 20 times slower then the host. Then it is going to be very slow. It would be like using a 100 Mhz CPU instead of a 2ghz.
Of course thats not to say it isn’t a useful tool or fun toy.
“All I meant by “WINE type” was that it runs binaries intended for a different platform than the host… as I am sure you are aware.”
But that is exactly what Wine doesn’t do. It allows you to run x86 (windows) apps on x86 Linux. No different platform.
Maybe there just trying something. Just like Emulating PSX and Sega and all those. It’s a hobby.
This is absolute anarcKy!!! DOGS AND CAT LIVING TOGETHER! PANDEMONIOUS MADNESS!!!
Please for the love of god kill this abomination, and free that poor OSX installation from having to endure another moment on that worthess X86 box. The MACOS running inside windoze is about the most abhorant and abombinably disgusting thing that I have ever seen.
You and pearpc should both be utterly and completely ashamed of yourselves for this travesty!!!
Look with less than 5% of the market and shrinking, Mac users should just face reality – they are sick misguided morons
I mean the consumers have made their choice, bith corporate and home users have clearly selected anything but Mac over years and years.
But the shite that Mac zealots come out with OMG…
I detest spending more than say $800 on a PC, and for that money I wouldn’t get anything in the modern Mac range (maybe an iPOD spare battery, and god knows I will need it!) Where as in x86 space, I can by a 64 bit chip, overcklock the crap out of it, and then upgrde the box for years to come
decent games are all for the PC (or consoles) but you never see a TV advert saying: “Coming SOON for Mac and PS2” cause nobody (ok 95% of people) gives a flying donkey cock about what happens to Mac.
Therefore – excellemt project, lets devour the last dignity that Mac may have (its camp, but pleasing looks) and use if for our own good…
Now that ought to stir up some more comments
Yet another useless open source software.