Ease off bashing AmigaONE X1000?

[SamuraiCrow]

I think AEON, as a company, is trying to get stuff sold.  They are following a dead-end track to the very end.  Once they get there I don't think there is a plan B to save their efforts.  I think Trevor Dickinson is making an honorable effort to try to push the plan A solution but there really isn't much of a market for PPC desktop computers as a whole nowadays.

When I found out that AmigaOS 4.x was going to stay PPC-only even though improved FPGA-based solutions started presenting themselves as a sequel to the Classic Amiga chipsets, I reverted to AmigaOS 3.x and started looking toward AROS as a solution for any next-generation style needs.  I especially value AROS 68k as a potential improvement over and above AmigaOS 3.x as I start adding features that the chipsets have that the OS never adequately supported.

My vote:  No.  Bashing the X1000 is justifiable given that the AmigaOS team will forever be understaffed due to being a commercial venture with a limited market.  Even the closed-source shareware route of MorphOS is of limited usefulness since it is PPC only and also understaffed (though MorphOS is faster than AmigaOS 4.x since MorphOS doesn't use interfaces in its shared libraries).

I think that the only hobby operating systems worth mentioning are AROS and Haiku.  AROS, because it is open-source, supports a number of processor architectures, and is largely self-supporting.  Also Haiku because it is multithreaded and pervasively object-oriented (not to mention largely open-source).  Both of those are fairly quick compared to modern commercial OSs on modern hardware.

[SamuraiCrow]

The path to perfomance in this day and age is parallel processing.  A dual-core CPU in this day and age is pathetic in the grand scheme of things (although I'm still running on a Core 2 Duo on my Mac Mini and a dual-core Centrino for my Windows 7 laptop).  In order to get the level of parallelism out of a CPU, Hyperion's going to need to trash most of the compatibility argument.  This renders their efforts rather futile since Amigans are the only link to the past (other than the Legend of Zelda title  :laugh1: ).

If they were going to do a clean break, they should have done it early on.  Personally, I think that an FPGAArcade can beat the snot out of an X1000 if only its programmers could quit writing code in single-threaded C and instead write it in Erlang and have a compiler render to custom cores using VHDL as an intermediate representation for the FPGA.

FPGAs, despite their expense, are massively more parallel than a CPU can ever hope to be.  They are also more flexible and efficient than a GPU since it is more general-purpose than those parallel-pipelined beasties.  The time for dedicated silicon for specific features is quickly drawing to a close at the hands of field programmable gate arrays.

[SamuraiCrow]

I've met Trevor at AmiWest.  He is an Amigaholic and an honest hard-working one at that!

But as for AROS or MorphOS or even AmigaOS to win, I think will be a hollow victory even if it happens.  I suspect with the ability to build custom cores for the FPGAArcade, parallel performance of the FPGA itself will eventually kill off the Von Naumann architectures altogether.  This includes ARM, AMD64, x86, PPC, and every other microprocessor known, even 68k.  For more information on high-level synthesis of FPGA cores, see these TCE screenshots and the associated Linux downloads.