Linked by Thom Holwerda on Mon 20th Dec 2010 19:56 UTC
Amiga & AROS A few moments ago, a woman appeared at my doorstep with a package for me. Since it was about 19:30, I was a little perplexed - this isn't a very usual time to be delivering packages. Also, this was a woman who I've seen walking around town before - she's a local. She handed over the package, and when I saw the note on top, I was even more intrigued: "From Trevor Dickinson, A-EON Technology".
E-mail Print r 24   · Read More · 40 Comment(s)
Thread beginning with comment 454345
To view parent comment, click here.
To read all comments associated with this story, please click here.
Member since:

On paper, it seems like a very good CPU. Does anyone know what the consequences will be for the Amiga OS? does this CPU offer greater functionality and performance than Intel CPUs? what is the advantage of this CPU?

It was a great cpu in its time (a few yeasr it would have attraced me stropngly) and today is still a nice one, ableit waaay to expensive and not really up to date.
The reason for OS4 to chose PA6T is plain and simply that it is a ppc and thus offers a big endian data type. And OS 4 is ppc/big endian only. A switch to x86/little endian for that OS is far from trivial if you want to keep backwards compability.
Diffrent to most other OSes OS4 shares structs between the OS and apps, if these dont match the same endianess you are in deep trouble. The same is true for MorphOS.
If the PA6T or any othe rppc is only a particuar choice, but generally the PA6T is a nice cpu, albeit not developed any longer and not my first choice (which Freescale 86xx or some QorIQ chip).

Reply Parent Score: 1

billt Member since:

IMHO PA6T is supoerior to Freescale's 86xx choices. PA6T is a G5 class 64bit core, 86xx is a G4 class 32bit core. QorIQ for the moment I think also is less desirable than PA6T, though they do finally have 64bit and seem to plan a return of Altivec, so I do very much look forward to news there for something I hope will be comparable to PA6T and have a future which PA6T does not.

PA6T has a good number of SERDES for PCI-Express and other things, with more configuration options than other PPCs, and until QorIQ adds a couple more things I think PA6T is still king of PPC, and however they managed to get some chips I think it's a fantastic choice when restricted to PowerPC.

Reply Parent Score: 1

TheGZeus Member since:

I'm no expert on the way these OSen work, so bear with me.

Would it be easier to port to say, Ultra/OpenSparc?
No royalties involved, you could run it on an FPGA board for fun...
It's big-endian...

Reply Parent Score: 2

axilmar Member since:

Diffrent to most other OSes OS4 shares structs between the OS and apps, if these dont match the same endianess you are in deep trouble. The same is true for MorphOS.

Don't they have the appropriate preprocessor definitions, macros and functions that would allow them to solve the endianess issue?

Reply Parent Score: 2

Carewolf Member since:

I believe x86 now has extensions to load and store big-endian. It could easily be made a compiler option to simply switch the endianness. Even without the a simple runtime option, the endian could still be configured in the compilers OS description file.

Reply Parent Score: 2