Yokemate of Keyboards
Posts: 4977 from 2009/1/28
From: Delaware, USA
Quote:In_Correct wrote:
Quote:takemehomegrandma wrote:
Quote:
Jim wrote:
Disadvantages?
Like compatibility with my current hardware, the possibility of boxing in legacy apps, etc?
Those "disadvantages"?
After the legacy cord has been cut, legacy apps will either face to be boxed-in or ported. This no matter if it's PPC or x86, the "only" difference being that the advantage upside for x86 is huge, and the advantages speaking for PPC is, what... that you already have a machine? Same for addressable memory, a feature like virtual memory makes so much more sense in a 64-bit memory space than a 32-bit, and the only thing speaking in favor would again be that you already have a 32-bit machine? And that was just the addressable memory space of it, then we have the rest of the 64-bit computing benefits...
Quote:
Yes, as I have said before, even if we fork a branch to X64, I like to see the developers consider those advantages in the PPC version.
A fork instead of a migration is perhaps plausible... But wouldn't the strengths of the PPC line be exactly the "legacy cord", the point in using it. The Amiga3.1 environment, API, endianness etc. Highest possible compatibility always being a priority. The possibility to run the Amiga legacy stuff native on MorphOS mixed with todays PPC MorphOS apps. Add SMP, 64-bit etc to this, and you will have destroyed this benefit and turned the PPC fork into a cripled wannabe of the so much better x86 fork, no x86 advantages and no "legacy" advantage either. Better let the PPC fork be untouched from "new stuff" IMHO, solid legacy support is what that fork is and will be about.
IMHO, The PPC Fork should reflect The X64 Fork as well as reflect The ARM Fork And the MIPS Fork.
I am aware that multiple forks are not possible.
But beginning with The PPC Fork, the advantage of "already having a machine" is reasonable. I am curious to see a MorphOS fully supported 64 bit m
ulti core PPC device.
Thanks, since we already support one 64bit system (the G5) and adoption of another has been announced (the X5000), I too would like to see some support of the features these machines allow.
And, no, this would NOT have to violate compatibility because it does not have to be carried out within Abox.
And there is nothing preventing code running outside of Abox from working with code running inside Abox.
So the breaking compatibility argument is not valid.
Certainly this is a more difficult course than a clean slate OS, and I would still support an X64 fork.
But really, even with a move to X64, does anyone think PPC MorphOS will just disappear?
With its current licensed base, even without further development I would expect to see our legacy systems still in use for many years.
So why not exploit features we already have in our current hardware?
My G5 has 6GB of unused address space and one unused processor, a PCIe G5 or the X5000 could offer up to three unused cores.
Let me run code outside Abox, and I can prove what could be done without ditching our legacy OS.
[ Edited by Jim 22.08.2015 - 12:08 ]"Never attribute to malice what can more readily explained by incompetence"
