Transitive Technologies Powers Rosetta

[MacRumors]

News.com confirms that Transitive Technologies is, in fact, the technology behind Apple's Rosetta PowerPC Emulator for their upcoming Intel-based Macintoshes.

Jobs reportedly confirmed Transitive's role in a New York Times interview, but in general, Apple has been very quiet about their Transitive's role in Rosetta. Of note, Jobs' keynote speech on Monday gave no mention to the startup.

It appears Transitive's technology can provide 60-80 percent performance of native software based on real world experience with SGI. Some analysts, however, have doubts about the performance promises.

First mention of Transitive came in July 2003. The most accurate and earliest rumor about Transitive's use by Apple came as a Page 2 news item in February 2005:

...there is evidence that Apple has had special internal seeds of Tiger which support [Transitive's technology] for the x86 platform. Beyond allowing Tiger to run on x86, perhaps more significantly is the potential to also allow existing Mac OS X applications to be run on the x86 (PC) platform without recompilation.

Apple, of course, is not offering Mac OS X for the PC, but instead offering Intel-based Macs.

 

[Peace]

News.com confirms that Transitive Technologies is, in fact, the technology behind Apple's Rosetta PowerPC Emulator for their upcoming Intel-based Macintoshes.

Jobs reportedly confirmed Transitive's role in a New York Times interview, but in general, Apple has been very quiet about their Transitive's role in Rosetta. Of note, Jobs' keynote speech on Monday gave no mention to the startup.

It appears Transitive's technology can provide 60-80 percent performance of native software based on real world experience with SGI. Some analysts, however, have doubts about the performance promises.

First mention of Transitive came in July 2003. The most accurate and earliest rumor about Transitive's use by Apple came as a Page 2 news item in February 2005:



Apple, of course, is not offering Mac OS X for the PC, but instead offering Intel-based Macs.

Man I was wondering about this.Now I know..

So we're seeing the Transitive at work in this "transition"

 

[savar]

So is QuickTransit/Rosetta a chip or a piece of software?

 

[corywoolf]

Man I was wondering about this.Now I know..

So we're seeing the Transitive at work in this "transition"

Old news... I already said this four days ago.

https://forums.macrumors.com/showthread.php?p=1501709#post1501709

 

[corywoolf]

I think most people that are regulars here already knew (or had a good bet on it) this was the way they would switch to intel.

 

[joeboy_45101]

I wonder what type of performance we will see after a year of further development and tweaking on the part of Apple and Transitive. We may see near-full performance using Rosetta when it ships.

One thing I have been thinking about is this, Steve said that they would probably have the first Intel based Macs ready to show and probably sell by WWDC 2006 but Leopard is not going to ship for some time after that. So, I am wondering what version of Mac OS X will be shipped on these Macs. Is much of the software needed for the Intel based Macs already present in Tiger? Or will Apple be shipping a new version of Tiger that can be installed on either the PowerPC platform or the Intel platform.

 

[RBilRamZ]

confused?

i could be wrong, but doesnt this seem like its only a minor hack away from being able to put OSX on any old PC. i dont know how to do it myself, but it seems like once the switch to intels is made, it wouldnt me to hard for someone to manipulate their PC to accept and run OSX, any thoughts?

 

[barneygumble]

i could be wrong, but doesnt this seem like its only a minor hack away from being able to put OSX on any old PC. i dont know how to do it myself, but it seems like once the switch to intels is made, it wouldnt me to hard for someone to manipulate their PC to accept and run OSX, any thoughts?

i seriosuly doubt that this will work, only a limited number of hardware configurations will work and apple will be the only ones supplying them

 

[fox2005]

i think macs running on intel will have a different chipset, something like altivec or equivalent or some other propietary chip that will add funtionality and make it "different" than normal wintel machines... it will also make difficult to run plain windows from the box on them...

 

[gwangung]

i could be wrong, but doesnt this seem like its only a minor hack away from being able to put OSX on any old PC. i dont know how to do it myself, but it seems like once the switch to intels is made, it wouldnt me to hard for someone to manipulate their PC to accept and run OSX, any thoughts?

There are some Darwin folks over on ars.technica involved in porting Darwin to other chipselts, and according to them, it's far from a minor hack.

Some people dismiss that, but I would think practical experience would trump web board BSing....

 

[Cybernanga]

I wonder what type of performance we will see after a year of further development and tweaking on the part of Apple and Transitive. We may see near-full performance using Rosetta when it ships.

The current "reported" performance of around 805% is pretty darn impressive, I sincerely doubt that they'll be able to tweak it much more that that.

One thing I have been thinking about is this, Steve said that they would probably have the first Intel based Macs ready to show and probably sell by WWDC 2006 but Leopard is not going to ship for some time after that. So, I am wondering what version of Mac OS X will be shipped on these Macs. Is much of the software needed for the Intel based Macs already present in Tiger? Or will Apple be shipping a new version of Tiger that can be installed on either the PowerPC platform or the Intel platform.

Steve also said that OS X has been "leading a double life". They've had it running on Intel for the last five years, so I figure most of the code is already in ther, and just needs to be "activated". Hopefully a Software Update (restart required, natch) will do the trick.

 

[Cybernanga]

So is QuickTransit/Rosetta a chip or a piece of software?

Rosetta is Software.

 

[bousozoku]

There are some Darwin folks over on ars.technica involved in porting Darwin to other chipselts, and according to them, it's far from a minor hack.

Some people dismiss that, but I would think practical experience would trump web board BSing....

Yes, if it was a minor hack, we would have had BeOS version 5 on Blue and White boxes around 1999.

It seems to me that we were talking here about Transitive months ago and their claims then seemed quite bold. Obviously, they've been able to accomplish quite a lot.

 

[mkrishnan]

There are some Darwin folks over on ars.technica involved in porting Darwin to other chipselts, and according to them, it's far from a minor hack.

Well...PearPC shows that it's possible in principle to get emulated MacOS running on an Intel box. PearPC doesn't require roms, does it? The only question is how far in the direction of a bootstrap loader can you get, from Linux or Win + PearPC, which isn't anything like the footprint of a bootstrap loader.

*And* how is it that Transitive is so fast? I was one of the people who didn't believe that such a piece of software could accomplish what it seems to. And I seem to be wrong!

But I do agree that PearPC seems to have nowhere near the (still limited) effectiveness of Rosetta.

 

[gwangung]

Well...PearPC shows that it's possible in principle to get emulated MacOS running on an Intel box. PearPC doesn't require roms, does it? The only question is how far in the direction of a bootstrap loader can you get, from Linux or Win + PearPC, which isn't anything like the footprint of a bootstrap loader. .

Hmmm....I may be mistaken, but emulating hardware may not be as hard as hacking into custom ASICs that Apple are sure to use and to use that in system work...

 

[mkrishnan]

Hmmm....I may be mistaken, but emulating hardware may not be as hard as hacking into custom ASICs that Apple are sure to use and to use that in system work...

Yeah, true.... Apple can definitely engineer a way that would rule out "hacks" that allow for OS X to natively run on Wintel hw. I thought the previous poster (RBilRamZ) had meant something more along the lines that, if Rosetta can emulate so efficiently, then it is possible in principle to develop another emulator, say PearPC, to the point where it also had this level of feasibility.... But I don't think, in hindsight, that was what was meant at all.

 

[Amdahl]

Sounds a bit like Java HotSpot

The process of translating code the first time it is run sounds a bit like modern Java virtual machines.

Do PPCs do self-modifying code? I bet that would really hurt performance.

 

[steeldrivingjon]

Benchmarking Rosetta

If anyone is going to try running XBench on an Intel Mac to test Rosetta performance, I would suggest doing the following:

Run the test twice, without quitting, and use the second set of results.

As the benchmark goes through the various tests, Rosetta will cache the translated binary. But it won't get to use the cache as would normally be the case, because a benchmark doesn't go back to repeat prior steps.

In the first run through the benchmark, Rosetta will have to do the translation to Intel, which slows things down even more. In the second run through the benchmark, Rosetta will be able to use the cached Intel code, which should result in better performance the second time around.

If you just run XBench once, and take the results, you won't get a realistic benchmark. of Rosetta. In real use, Rosetta will have a PowerPC app's cached binary around, and most of the time it won't have to re-translate. So it'll perform better.

 

[Sabenth]

Well...PearPC shows that it's possible in principle to get emulated MacOS running on an Intel box. PearPC doesn't require roms, does it? The only question is how far in the direction of a bootstrap loader can you get, from Linux or Win + PearPC, which isn't anything like the footprint of a bootstrap loader.

*And* how is it that Transitive is so fast? I was one of the people who didn't believe that such a piece of software could accomplish what it seems to. And I seem to be wrong!

But I do agree that PearPC seems to have nowhere near the (still limited) effectiveness of Rosetta.

Limited how can that be limited iam not a developer iam no code exspert from what i saw this software is very very very dam effective .. Sorry not having a go very early morning here havent had me java yet ...

 

[SiliconAddict]

The current "reported" performance of around 805% is pretty darn impressive, I sincerely doubt that they'll be able to tweak it much more that that.

Don't know about you guys but I'd be happy at 805% performance over native PPC speed. Isn't that the point where the app opens before you are done clicking?

 

[Omni Geno]

If anyone is going to try running XBench on an Intel Mac to test Rosetta performance, I would suggest doing the following:

Run the test twice, without quitting, and use the second set of results.

As the benchmark goes through the various tests, Rosetta will cache the translated binary. But it won't get to use the cache as would normally be the case, because a benchmark doesn't go back to repeat prior steps.

In the first run through the benchmark, Rosetta will have to do the translation to Intel, which slows things down even more. In the second run through the benchmark, Rosetta will be able to use the cached Intel code, which should result in better performance the second time around.

If you just run XBench once, and take the results, you won't get a realistic benchmark. of Rosetta. In real use, Rosetta will have a PowerPC app's cached binary around, and most of the time it won't have to re-translate. So it'll perform better.

I'd really like to see some of these results.

 

[swissmann]

And so many of us (including myself) thought that Transitive was vaporware.

 

[Davito]

I am afraid that when as all the lines are transported to Intel chips, 3rd party developers will soon stop compiling for PPC. Therefor I would like to see a 'Reverse-Rosetta' to keep my G5 running new software for some more years!

 

[steeldrivingjon]

I am afraid that when as all the lines are transported to Intel chips, 3rd party developers will soon stop compiling for PPC. Therefor I would like to see a 'Reverse-Rosetta' to keep my G5 running new software for some more years!

Why would they give up all those potential sales?

 

[mkrishnan]

Limited how can that be limited iam not a developer iam no code exspert from what i saw this software is very very very dam effective .. Sorry not having a go very early morning here havent had me java yet ...

Limited in what it emulates, not how well it emulates. According to the Apple developer software, it doesn't run any software that requires a G4/G5, it doesn't translate AltiVec, it doesn't run Classic on OSX/Intel (not that I care) and there are a number of other limitations. Not bad, all in all, but it definitely isn't a 100% software emulation of everything that you could do on a real G4 or G5....

But don't get me wrong. As I said, I'm duly impressed by what I've seen so far.