Somebody finally made a Mac emulator, well one for PowerPC Macs that is. They have been around since 1994, but until now, we could only emulate the older 68k-based Macs on PC.
It is called Pear (which I nominate worst pun of 2004) and you can find it here: http://pearpc.sourceforge.net/
So if you want to see OS X in its full glory, and see what you have been missing by using an x86 and running Windows, or even worse, linux... now you can! :)
Oh, and if you use the JIT-compiled version, you only get 1/40th of your CPU's performance in the emulated PowerPC, instead of the 1/500th for the non-JIT one.
So with my 1533 MHz PC, I will get a nice 38 MHz Mac, nice.
It is called Pear (which I nominate worst pun of 2004) and you can find it here: http://pearpc.sourceforge.net/
So if you want to see OS X in its full glory, and see what you have been missing by using an x86 and running Windows, or even worse, linux... now you can! :)
Oh, and if you use the JIT-compiled version, you only get 1/40th of your CPU's performance in the emulated PowerPC, instead of the 1/500th for the non-JIT one.
So with my 1533 MHz PC, I will get a nice 38 MHz Mac, nice.
Oh, and if you use the JIT-compiled version, you only get 1/40th of your CPU's performance in the emulated PowerPC, instead of the 1/500th for the non-JIT one.
So with my 1533 MHz PC, I will get a nice 38 MHz Mac, nice.
There are a lot of people that whant see the mac emulation :D (ok I, bro and a friend ;) ), this is nice
Have a nice day or night.
Have a nice day or night.
Sounds like they need ASM programmers. ;)
Or maybe the PowerPC chip has many addressing modes, decoding this addressing mode block, man crazy.
x86asm, just requires a good dispatcher. For example, table-less:
mov eax, OFFSET Routines
mov ah, [esi]
inc esi
jmp eax
ALIGN 256*256
Routines:
Routine_00:
retn
ALIGN 256
Routine_01:
retn
...
ALIGN 256
Routine_FF:
retnActually the PowerPC chip doesn't have many addressing modes at all, being a RISC design.
Only the ld/st instructions can access memory, and there are only a few addressing options. I believe only register, indexed and pre-increment are supported.
I've programmed asm on them years ago. Great architecture, but hell to emulate on x86, because the instructions are more powerful (some operations require 3 or 4 x86 ones), have 3 operands (separate destination) and there are 32 gp registers and 32 FPU registers.
Only the ld/st instructions can access memory, and there are only a few addressing options. I believe only register, indexed and pre-increment are supported.
I've programmed asm on them years ago. Great architecture, but hell to emulate on x86, because the instructions are more powerful (some operations require 3 or 4 x86 ones), have 3 operands (separate destination) and there are 32 gp registers and 32 FPU registers.
In that case their JIT method is probable the best way to go. Direct emulation should be better than 1/500th though.
In that case their JIT method is probable the best way to go. Direct emulation should be better than 1/500th though.
definitely, I am calling functions to read bytes in and write bytes out of/to RAM and my Z80 CPU emul can beat that (possibly because PowerPC CPU is more complex?, even tho being RISC).My emu is interpreter based, it would be nice to see how dynamic recompliers work.
I know that RISC is more faster than CISC , so what is an acceptable/average speed for a Mac ? Yes , emulating will not faster than a real Mac but whether it can acceptable ?
JIT is always the best way to go. And personally I don't really care how fast their interpreter is, since the JIT is faster anyway. But the JIT is not very good either, I think. If you look at VirtualPC (x86 emulator on Mac), it gets much better results. Same for the 68k JIT in WinUAE.
As for what an acceptable speed is for a Mac, that depends on the user, the programs he runs, and the standards he has for performance. But if you only get 1/40th of the native speed, I guess it's just too slow to be useful at all.
But this is just the first version, the project is brand new. They should be able to improve lots, let's hope that they do (help them if you like). With a decent JIT it would not be unthinkable that they will eventually get 1/5th performance, or better. Then it should get interesting.
As for what an acceptable speed is for a Mac, that depends on the user, the programs he runs, and the standards he has for performance. But if you only get 1/40th of the native speed, I guess it's just too slow to be useful at all.
But this is just the first version, the project is brand new. They should be able to improve lots, let's hope that they do (help them if you like). With a decent JIT it would not be unthinkable that they will eventually get 1/5th performance, or better. Then it should get interesting.
Oh, and if you use the JIT-compiled version, you only get 1/40th of your CPU's performance in the emulated PowerPC, instead of the 1/500th for the non-JIT one.
So with my 1533 MHz PC, I will get a nice 38 MHz Mac, nice.
So, to get a 100 MHz Mac, you need a PC of 4 GHz which is not available yet. :)
4GHz is available - if you got liquid nitrogen ;)
And there's 3.4GHz P4's anyway, and MHz isn't everything anyway...
And there's 3.4GHz P4's anyway, and MHz isn't everything anyway...
f0dder,
I would prefer solid carbondioxide instead of liquid nitrogen. :)
Do your remember for the scene in Terminator II - The judgement day , T1000 breaking into pieces after a liquid nitrogen bath? :)
I would prefer solid carbondioxide instead of liquid nitrogen. :)
Do your remember for the scene in Terminator II - The judgement day , T1000 breaking into pieces after a liquid nitrogen bath? :)