[News] Transmeta don't plan 68k support | ANN.lu |
Posted on 14-Feb-2000 20:02 GMT by Christian Kemp | 33 comments View flat View list |
Black Hand says he repeatedly mailed Transmeta to enquire about 68k emulation support, or the posibility to access and modifiy the morphing code. Here's what Frank Priscaro replied: Supporting the Amiga was something that we considered early on, but had to drop because all of our resources needed to be focused on the Linux and x86 markets. We have no plans at this time to support the 68K family of processors.
|
|
List of all comments to this article |
Transmeta don't plan 68k support : Comment 33 of 33 | ANN.lu |
Posted by Mario Saitti on 16-Feb-2000 23:00 GMT | In reply to Comment 32 (Shaun Murray): > 1) Everyone is missing the point of the crusoe technology. It doesn't just
> emulate a CPU architecture, it can emulate the whole of the PC's architecture > (or any other for that matter). Ie. do what UAE does in hardware.
Not missed, just not mentioned. Also it does not stop at UAE's limitations, it is closer to FX32 than UAE. For the benefit of others here, (I will try to make this as simple as possible) there are numerous reasons for using this combined hard/soft methodology. Rather than just use hardware assists to speed up emulation(The easiest to grasp example is Crusoe's translation cache which stores the already translated "commands"), it also uses software trickery to alter the program running itself optimising it as it runs, this is done on numerous levels one such level being dynamic recompilation(Basically rewriting/optimising the code) as FX32 on the Alpha.
> 2) I'm surprised they haven't got 68K on the horizon as crusoe would maybe
> make a good processor to get the PalmOS people off low powered Dragonball
> chips which are pretty much 68K cores with added memory controllers, LCD
> panel controllers etc. The problem is that so far the power requirements of
> the chip are still a little on the high side for PDA computing.
There are numerous reasons why it might not be on the horizon. Firstly the complexity. They can't just swap architecture. Secondly the time frame. The very fact that it took so long to produce an X86 emulation suggests it could be at least 2 years for 68k.
VLIW compilers are one the most complex aspects of computing I can think of, it is the slowest aspect of computing in regards to progress iirc. Now knowing that, imagine the burden of emulation piled atop of an already complex task, then add dynamic recompilation to the equation and look at the task before them.
They have to evaluate whether the development costs can be recouped. It appears that they already have. |
|
|
- User Menu
-
- About ANN archives
- The ANN archives is powered by #AmigaZeux. It was updated daily (news last: 22-Oct-2004; comments last: 18-May-2005).
ANN.lu was created, previously owned and maintained by Christian Kemp, www.ckemp.com.
- Contribute
- Not possible at this time!
- Search ANN archives
- Advanced search
- Hosting
- ANN.lu was hosted by Dreamhost. Sign up through this link, mention "ckemp" as referrer and he will get a 10% commission on any account you purchase.
Please show your appreciation for any past, present and future work on ANN.lu by making a contribution via PayPal.
|