So... The new Mac Office is out, but where the HECK is the Windows Emulator that is capable of running on a G5. They have had over one year to make this!!! Shit microsoft is slow at updating their software!
FIrst of all, this isn't Microsoft's fault. I'd wager a bet and say if they hadn't bought Connectix this would be taking even longer. VPC makes use of a lot of hand-written assembly that doesn't work on the G5. Have you ever written in assembly? Not too hard but it's slow as hell to get anything done.
Anyway, in the meantime you can try bochs or QEMU if you have a G5.
FIrst of all, this isn't Microsoft's fault. I'd wager a bet and say if they hadn't bought Connectix this would be taking even longer. VPC makes use of a lot of hand-written assembly that doesn't work on the G5. Have you ever written in assembly? Not too hard but it's slow as hell to get anything done.
Anyway, in the meantime you can try bochs or QEMU if you have a G5.
That's crap. Virtual PC is actually a very small application.
Well, VPC is a very small app, but assembly is also very small, that is why in the olden days everything was written in assembly. The limited memory footprints of machines mandated assembly. It can also be fast, and it also exposes features that high level languages have no concept of. In the case of VPC I believe this is the endianess setting.
Fixing the endianess issue may have been harder than they thought. VPC is unlikely to include an infrastructure for swapping bytes, and there are quite a few places to get it right. Adding all of that could easily have introduced some nasty bugs. In particular, if the rumor that they are trying to add native video card support is true, interfacing with real big endian hardware from little endian software will be tricky.
Another factor is that doing endianess in software is slow, and the result may be considered unmarketable ( eg: new G5 is slower than old G4 ).
Other systems have managed it, so Im sure it can be done, not so sure about the speed.
Well, VPC is a very small app, but assembly is also very small, that is why in the olden days everything was written in assembly.
Assembly was small when ISA's were CISC. Assembly is not so small with RISC because each instruction does a lot less. VPC is not small because it is written in assembly, it is small because interpreting x86 code is not a particularly big job.
Anyway, in the meantime you can try bochs or QEMU if you have a G5.
I can vouch for QEMU working on a G5. It took a certain degree of playing around and looking at UNIX commands I didn't quite understand (but which actually start to make a lot of logical sense after awhile), but I have got Win 98 and Win XP running on a SP G5 1.8. Not necessarily useful yet (no USB, but I can go on-line, at least with Win 98 ), but a very cool proof of concept.
Assembly was small when ISA's were CISC. Assembly is not so small with RISC because each instruction does a lot less. VPC is not small because it is written in assembly, it is small because interpreting x86 code is not a particularly big job.
You might be surprised at the complexity of the powerpc ISA. CISC and RISC are really philosophies, and not descriptions of what the ISA is. This is one reason why x86 has been able to keep up with risc designs, it is, in essence, a risc design, but that isnt reflected in its ISA.
Any way, the comment i was replying to implied that because VPC is small it should be easy to modify. What I was saying is that small code can still be complex, in particular, assembly can generate very small code, but have a significant number of lines, and potential for complexity.
Months ago, I read that VPC 7 will be revolutionary in that it will, for the first time, access the 3D graphics card and run graphics through it rather than straight software-- thus, Mac gaming will be advanced considerably.
Can anyone confirm this?
Can anyone comment on whether or not such a program, say running on a Dual G5 2.5 using VPC 7, will be able to run a Windows-only game with good speed, or will there be massive degradation anyway? Or will the fast 2.5 make up the difference?
You might be surprised at the complexity of the powerpc ISA. CISC and RISC are really philosophies, and not descriptions of what the ISA is. This is one reason why x86 has been able to keep up with risc designs, it is, in essence, a risc design, but that isnt reflected in its ISA.
Any way, the comment i was replying to implied that because VPC is small it should be easy to modify. What I was saying is that small code can still be complex, in particular, assembly can generate very small code, but have a significant number of lines, and potential for complexity.
Despite what Intel and its apologists say about the latest Intel x86 processors, they are not RISC. By making the assertion that x86 is RISC inside, Intel and its supporters implicitly admit that RISC is superior. Behind the decoder, all processors have elements of RISC and always have. Not only is the x86 not RISC, it is not even good CISC. For a good CISC design, take a look at the old National Semiconductor NS 32xxx family. Its ISA was so well-designed that there was almost a one-to-one relationship between a program line of a high-level language and the equivalent logic in assembly.
That would be freaking awesome but I doubt MS will be that cool. All the people that say it can't be done though are full of crap. It's already been done. They used to do it in VPC 3 or 4 for 3dfx cards. So it can be done. And with OS X's especially Tiger's graphics tech it's probably easier. Since most games these days are mostly GPU and since CPUs are getting real fast I imagine if VPC could use native graphics, than playing PC games in it might actually be possible. Then people might stop porting! Hmm that's a slippery slope...
That would be freaking awesome but I doubt MS will be that cool. All the people that say it can't be done though are full of crap. It's already been done. They used to do it in VPC 3 or 4 for 3dfx cards. So it can be done. And with OS X's especially Tiger's graphics tech it's probably easier. Since most games these days are mostly GPU and since CPUs are getting real fast I imagine if VPC could use native graphics, than playing PC games in it might actually be possible. Then people might stop porting! Hmm that's a slippery slope...
Wouldn't that be great? Rather than waiting like two years for a game to be ported to pick them right off the shelf? You'd think by now that with G5 speeds up at Dual 2.5, that they'd be able to crank out a decent PC Emu.
"What I've got here is a thousand monkeys on a thousand typewriters, writing the world's greatest book.."
I hope it will use native video card support, itd be a Godsend, id even be willing to throw down hundreds(or even the price of a new pc) for a card like apple used to sell (DOS Compatibility Card; it had its own REAL pentium processor with its own ram).
A Demo at Apple Expo Paris sounds likely to me. Windows XP SP2 is ready to ship and Amazon.com now has listed Sept. 13 as the date Office 2004 Pro (including Virtual PC 7) will be available. This stuff is important to MS, because the Xbox 2 is a G5.
.... This stuff is important to MS, because the Xbox 2 is a G5.
Nonsense. The Xbox 2 does not run any version of MacOS X 10. The only thing that the Mac and Xbox 2 have in common is that they both use siblings in the same processor family.
Comments
Anyway, in the meantime you can try bochs or QEMU if you have a G5.
Originally posted by Ra
FIrst of all, this isn't Microsoft's fault. I'd wager a bet and say if they hadn't bought Connectix this would be taking even longer. VPC makes use of a lot of hand-written assembly that doesn't work on the G5. Have you ever written in assembly? Not too hard but it's slow as hell to get anything done.
Anyway, in the meantime you can try bochs or QEMU if you have a G5.
Fixing the endianess issue may have been harder than they thought. VPC is unlikely to include an infrastructure for swapping bytes, and there are quite a few places to get it right. Adding all of that could easily have introduced some nasty bugs. In particular, if the rumor that they are trying to add native video card support is true, interfacing with real big endian hardware from little endian software will be tricky.
Another factor is that doing endianess in software is slow, and the result may be considered unmarketable ( eg: new G5 is slower than old G4 ).
Other systems have managed it, so Im sure it can be done, not so sure about the speed.
Originally posted by mmmpie
Well, VPC is a very small app, but assembly is also very small, that is why in the olden days everything was written in assembly.
Assembly was small when ISA's were CISC. Assembly is not so small with RISC because each instruction does a lot less. VPC is not small because it is written in assembly, it is small because interpreting x86 code is not a particularly big job.
WHATEVER!
All I care about is having the Windows Emulator run on my G5. When the heck are they going to release this product? Anyone know?
Originally posted by Ra
Anyway, in the meantime you can try bochs or QEMU if you have a G5.
I can vouch for QEMU working on a G5. It took a certain degree of playing around and looking at UNIX commands I didn't quite understand (but which actually start to make a lot of logical sense after awhile), but I have got Win 98 and Win XP running on a SP G5 1.8. Not necessarily useful yet (no USB, but I can go on-line, at least with Win 98 ), but a very cool proof of concept.
Originally posted by KCOM2006
Small... Big... Assembly Code...
WHATEVER!
All I care about is having the Windows Emulator run on my G5. When the heck are they going to release this product? Anyone know?
Microsoft has already said that it will release VPC 7 after it releases Win XP SP2. That should happen real soon now.
Originally posted by Mr. Me
Assembly was small when ISA's were CISC. Assembly is not so small with RISC because each instruction does a lot less. VPC is not small because it is written in assembly, it is small because interpreting x86 code is not a particularly big job.
You might be surprised at the complexity of the powerpc ISA. CISC and RISC are really philosophies, and not descriptions of what the ISA is. This is one reason why x86 has been able to keep up with risc designs, it is, in essence, a risc design, but that isnt reflected in its ISA.
Any way, the comment i was replying to implied that because VPC is small it should be easy to modify. What I was saying is that small code can still be complex, in particular, assembly can generate very small code, but have a significant number of lines, and potential for complexity.
Can anyone confirm this?
Can anyone comment on whether or not such a program, say running on a Dual G5 2.5 using VPC 7, will be able to run a Windows-only game with good speed, or will there be massive degradation anyway? Or will the fast 2.5 make up the difference?
Comments?
Originally posted by mmmpie
You might be surprised at the complexity of the powerpc ISA. CISC and RISC are really philosophies, and not descriptions of what the ISA is. This is one reason why x86 has been able to keep up with risc designs, it is, in essence, a risc design, but that isnt reflected in its ISA.
Any way, the comment i was replying to implied that because VPC is small it should be easy to modify. What I was saying is that small code can still be complex, in particular, assembly can generate very small code, but have a significant number of lines, and potential for complexity.
Despite what Intel and its apologists say about the latest Intel x86 processors, they are not RISC. By making the assertion that x86 is RISC inside, Intel and its supporters implicitly admit that RISC is superior. Behind the decoder, all processors have elements of RISC and always have. Not only is the x86 not RISC, it is not even good CISC. For a good CISC design, take a look at the old National Semiconductor NS 32xxx family. Its ISA was so well-designed that there was almost a one-to-one relationship between a program line of a high-level language and the equivalent logic in assembly.
Originally posted by Aquatic
That would be freaking awesome but I doubt MS will be that cool. All the people that say it can't be done though are full of crap. It's already been done. They used to do it in VPC 3 or 4 for 3dfx cards. So it can be done. And with OS X's especially Tiger's graphics tech it's probably easier. Since most games these days are mostly GPU and since CPUs are getting real fast I imagine if VPC could use native graphics, than playing PC games in it might actually be possible. Then people might stop porting! Hmm that's a slippery slope...
Wouldn't that be great? Rather than waiting like two years for a game to be ported to pick them right off the shelf? You'd think by now that with G5 speeds up at Dual 2.5, that they'd be able to crank out a decent PC Emu.
"What I've got here is a thousand monkeys on a thousand typewriters, writing the world's greatest book.."
I hate this PC so much
http://www.microsoft.com/mac/product...pid=comingsoon
Hopefully, a demo may be available at Apple Expo in Paris.
Originally posted by Franck
Microsoft says: Coming Soon !!
http://www.microsoft.com/mac/product...pid=comingsoon
Hopefully, a demo may be available at Apple Expo in Paris.
In other news, the Allies win World War II!
Originally posted by Maddan
.... This stuff is important to MS, because the Xbox 2 is a G5.
Nonsense. The Xbox 2 does not run any version of MacOS X 10. The only thing that the Mac and Xbox 2 have in common is that they both use siblings in the same processor family.