the_mole1314 said:So, when's Microsoft going under, since you can get Windows on a Mac...
So, do you want to figure out what is wrong w/that statement by yourself or would you like some help?
Lethal
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
PearPC - PowerPC/Mac OS X Emulation on a PC
- Thread starter MacBytes
- Start date
- Sort by reaction score
So, do you want to figure out what is wrong w/that statement by yourself or would you like some help?
Lethal
I see no point in buying a mac anymore if you happen to be a mac gamer. This is amazing!
You can use a PC to game and MacOSX for everything else! Sweet.
I wonder if we'll see an legal action from Apple against PearPC?
You can use a PC to game and MacOSX for everything else! Sweet.
I wonder if we'll see an legal action from Apple against PearPC?
i dont think most average users will go out and buy this and a copy of osx , its mainly for mac users who need to run stuff on windows machines. thats my take
markoibook said:
That used to be true, the Toolbox was in ROM. But with the iMac generation they moved the Toolbox to RAM and OSX does not rely on the Toolbox at all. Classic is a ROM-in-RAM machine. There were even some tools to load a real ROM into RAM for better OS9 performance.
There is an Open Firmware boot ROM, but it's, well, Open so that shouldn't be an issue.
With regard to the speed issue, I wouldn't get too comfortable with the Mac performance advantage for the causal user. I use a 350MHz G3 for alot of my work and it runs Panther just fine, even though the G5 2.0 is something like 20 times faster than it. That means an emulator can have a 20X performance hit and still have an OK OSX performance experience. The trouble with Virtual PC on Mac is that Windows XP runs like a dog on a 350MHz Pentium II, so there are some comparative advantages for OSX emulators.
bryanc said:
Doubtful. If a person is smart enough to know how to install and config the emulator part and then knows enough about installing an OS they "should" (Note the quotes.) be aware of how an emulator works.
I wouldn't think Apple is worried about this. I cannot imagine an emulator ever running FCP or DVD SP at acceptable speeds, and lets face it, those are the programs that sell the high end PowerMacs.
I think there is more a need for a good x86 emulator on the PPC platform than vice versa. The majority of Mac users I know would like to be able to run Windows programs occasionally, but not many Windows users are desperate to run iPhoto on their Dell.
I think there is more a need for a good x86 emulator on the PPC platform than vice versa. The majority of Mac users I know would like to be able to run Windows programs occasionally, but not many Windows users are desperate to run iPhoto on their Dell.
benpatient said:
You should note that WINE stands for "Wine is Not an Emulator." It does not emulate any hardware, it does not "boot" at all, and it is (quite obviously) not an emulator.
bryanc said:
if you know enough 1) that there are OSes other than windows and 2) to try to install this program (and make it work) on a x86 box, then you certainly know enough to realize it's an emulation and the speed isn't representative.
i think this is not much more than a gimmick - it's just a cool thing to try and implement. but i highly doubt that it will have much of an effect on the sale of Macs or OS X, even if it's made faster.
edit: siliconaddict - you beat me to it.
markoibook said:
This is no longer true, in fact Mac-on-Linux already allows you to run OS X on non-Apple hardware (PPC only).
But as the FAQ points out, this is technically illegal:
Alternative approach.
If Darwin is already running on x86, why cant people hack Darwin to trick Mac OS X into thinking its running on a PPC?
much less work that way.
The guts of the OS has already been ported. If people can 'convince' the OS that the chip is the same, everything should work fine.
Of course, its much easier said than done, and I have no clue how to do it.
If Darwin is already running on x86, why cant people hack Darwin to trick Mac OS X into thinking its running on a PPC?
much less work that way.
The guts of the OS has already been ported. If people can 'convince' the OS that the chip is the same, everything should work fine.
Of course, its much easier said than done, and I have no clue how to do it.
This is all very nice but where is Virtual PC 7(?) for the G5 that one rumour mill said would be out by February?
This is cool and about freaking time. A couple problems though. IMHO there are certain pro apps that barely run fast enough on OS X with a G4. Would you really want to run any iApp or god help you Photoshop or FCE on an emulator? I'm not a fan of the PPC architecture but I think its best to stick with the architecture that the OS was designed for. Unless you are a switcher and NEED to emulate it.
That's the only use for this, that I can see. Anyone who is switching to a PC. (Like it or not guys it DOES happen occasionally.) This would allow them to access any key software that is only available for the Mac.
This is a double-edged sword.
Its a good thing in that it could allow users who want to play with OS X (Ewww I just realized that means virus writers too.) the freedom to try it out and see that "hey Apple might be onto something here".
But also it has the ability to further tempt anyone who is eyeing the competition but doesn't want to give up a key app.
Don't know what to think of this.
That's the only use for this, that I can see. Anyone who is switching to a PC. (Like it or not guys it DOES happen occasionally.) This would allow them to access any key software that is only available for the Mac.
This is a double-edged sword.
Its a good thing in that it could allow users who want to play with OS X (Ewww I just realized that means virus writers too.) the freedom to try it out and see that "hey Apple might be onto something here".
But also it has the ability to further tempt anyone who is eyeing the competition but doesn't want to give up a key app.
Don't know what to think of this.
Spk1 said:
No one ever said Feb. It was announced to be released first half of 04. Which means no complaining until July.
Kinda off topic but this all reminds me of the guy that is running over 60 different operating systemss on his 17' PowerBook.
http://www.kernelthread.com/mac/vpc/
Waayyy too much time on his hands:
http://www.kernelthread.com/mac/vpc/
Waayyy too much time on his hands:
"if you're a mac gamer you have no more reason to buy a mac" AH ha ha, thats hilarious. Half of the people here say that it will never be fast enough to do any mac related work, and they say everybody should know that. And yet the other half says they're gonna jump ship from mac to PC but still run OSX. The people who say it won't be viable are correct. The reason? Well its the same reason we can't "trick" OSX into thinking it is on PPC hardware. To make an analogy, powered wheels drive a car, but you can't slap powered wheels on an airplane and expect it to fly. We can't run OSX on a PC because PC's don't have the same CPU in it. Meanwhile, this emulator doesn't just ALLOW mac OSX to run on x86, it is its own VirtualPPC. This means that it can run any PPC instruction software (maybe), but it will never run it anywhere near as fast as it does on a native Mac.
bree said:Alternative approach.
If Darwin is already running on x86, why cant people hack Darwin to trick Mac OS X into thinking its running on a PPC?
much less work that way.
The guts of the OS has already been ported. If people can 'convince' the OS that the chip is the same, everything should work fine.
Of course, its much easier said than done, and I have no clue how to do it.
Doesn't work that way. You can get darwin to run on a X86 platform but Apple's code, API's, etc are so tied to the PPC that it would require a rewrite of the OS along with every other app created for the OS. This is where the emulator is necessary.
OK, basic computer architecture lesson. Everyone knows (or should know) that when it comes to memory RAM is fast and Hard Drives are slow right? Well it goes beyond that you actually have a couple more layers. Going from slowest to fastest this is what you get:
Hard Drives
RAM
Cache
Registers
The registers are the fastests parts of computer memory, built right into the processor. These are the basic building blocks of a computer. Anything and everything that gets processed goes through the registers.
The PowerPC has 32 General Purpose Registers. This is where the bulk of the processing gets done. The Intel architecture has 8. Only 8. That means that any time you access more than 8 registers on the PPC side you'd have to start swapping information in and out of memory and that is going to be a massive bottleneck.
Think of it this way. Imagine you are at a library doing some research. You find yourself a nice table and start gathering the books you think you'll need. Well you can only fit so many books on the table at a time. Lets 8. You get the 8 books you think you'll need and start working. You find that you need some new info. So you take the least used book(s) and replace them with some new books. Now imagine you only have room for 2 books. (1/4 the space, just like the proccessors). You are going to have to get up and do a LOT more book swapping. This isn't just a 1:4 slow down ratio either. Since you have to swap more often you are spending alot more time in a very time-consuming process. The same is true with processors.
Now its true that the Intel architecture has some specialized registers which is why it doesn't run slow as molasses, but those aren't as large, so the PowerPC code won't be able to use them, the instructions wouldn't fit. So any emulator is going to have to either find a way to recast instructions without losing data (which means storing it on disk or at best in ram) or do a lot of register juggling.
Windows/Linux is fine because it is desgined to use those smaller registers for tasks, but the PPC architecture assumes that all registers (with a few exceptions) are 32bit and that there are 32 of them.
Sorry but this is never going to be a viable emulation standpoint even if they could get the instruction set translated. Its a hardware issue.
Hard Drives
RAM
Cache
Registers
The registers are the fastests parts of computer memory, built right into the processor. These are the basic building blocks of a computer. Anything and everything that gets processed goes through the registers.
The PowerPC has 32 General Purpose Registers. This is where the bulk of the processing gets done. The Intel architecture has 8. Only 8. That means that any time you access more than 8 registers on the PPC side you'd have to start swapping information in and out of memory and that is going to be a massive bottleneck.
Think of it this way. Imagine you are at a library doing some research. You find yourself a nice table and start gathering the books you think you'll need. Well you can only fit so many books on the table at a time. Lets 8. You get the 8 books you think you'll need and start working. You find that you need some new info. So you take the least used book(s) and replace them with some new books. Now imagine you only have room for 2 books. (1/4 the space, just like the proccessors). You are going to have to get up and do a LOT more book swapping. This isn't just a 1:4 slow down ratio either. Since you have to swap more often you are spending alot more time in a very time-consuming process. The same is true with processors.
Now its true that the Intel architecture has some specialized registers which is why it doesn't run slow as molasses, but those aren't as large, so the PowerPC code won't be able to use them, the instructions wouldn't fit. So any emulator is going to have to either find a way to recast instructions without losing data (which means storing it on disk or at best in ram) or do a lot of register juggling.
Windows/Linux is fine because it is desgined to use those smaller registers for tasks, but the PPC architecture assumes that all registers (with a few exceptions) are 32bit and that there are 32 of them.
Sorry but this is never going to be a viable emulation standpoint even if they could get the instruction set translated. Its a hardware issue.
Surprised no-one has suggested using PearPC to add PC hardware to an Xgrid cluster.
It should be fine in its current state to run simple (but time consuming) calculations. Maybe fourier transforms for SETI, etc, etc...
And if you have a roomful of idle PCs (Many Universities have numerous roomfuls every night), 1/40th or even 1/500th of their real power is still extra power. And of course its bound to get a bit faster.
It should be fine in its current state to run simple (but time consuming) calculations. Maybe fourier transforms for SETI, etc, etc...
And if you have a roomful of idle PCs (Many Universities have numerous roomfuls every night), 1/40th or even 1/500th of their real power is still extra power. And of course its bound to get a bit faster.
Will be interesting to watch this project... Can they ever optimize it enough to be useful is the big question? It's (relatively) easy to emulate old processors with the latest high power CPUs, but quite a different story to emulate something like a G4. 500 times slower is awfully slow right now.
invaLPsion said:
Er....have you ever used CPU emulation software before? With a fast enough computer you only get irritated at the speed. With a subpar computer with a an older CPU you get qwerty face from slamming your head into the keyboard.
I wonder if we'll see an legal action from Apple against PearPC?
Not possible. Since there is nothing illegal about creating an emulator. Hell if it did come down to a lawsuit they would throw Virtual PC back in the face of Apple.
For more information see the slashdot.org thread on this subject
For more information see the slashdot.org thread on this subject
http://developers.slashdot.org/arti...thread&tid=107&tid=162&tid=185&tid=187&tid=99
For more information see the slashdot.org thread on this subject
http://developers.slashdot.org/arti...thread&tid=107&tid=162&tid=185&tid=187&tid=99