UltimatePPC rumor

[Iggy]

Isn't this the PPC accelerator that uses a CPU that is binary incompatible with CPUs used in existing accelerators and motherboards?

Yes, they claim they are using a Freescale QoriQ P1013, and that is e500v2 based and features a very non-standard floating point unit.
The MorphOS development team has been quoted as saying that no other e500 core except the e500mc would be compatible with that OS.
And it is quite a step from the 603 cores used in Amiga accelerators.
It will be interesting to see how this works.

MorphOS support is very unlikely, as beyond the FPU issue, that would still require RTG.

But Amiga OS4 might be made to work on the device.

[Iggy]

I doubt you find usable x86 SoC that is cheaper than PPC.

I bet ultimateppc people have changed P10xx chip to a T10xx. (eur40...50 & 1.4Ghz).

Actually, an X86 SoC would be a lot cheaper, but a PPC would make more sense.
And you are right that a T1022 or T1042 would be a likely choice. It would make OS4 or MorphOS support easier as those use the same e5500 core that is used in the processor used in the X5000.
Legacy PPC support? Is that really that important? After all, you are talking about a small handful of applications.
Running OS3.1 apps under OS4.1 with real hardware for apps that directly access specific locations, now that could get interesting.

[Iggy]

@ Kimmok,

Yes, X86 IS a lot cheaper.
I could give you examples using desktop cpus, but they require an additional chipset.
So...here is a four core AMD Soc that runs at 2.4 GHz AND has a built in GPU.
http://www.semiconductorstore.com/cart/pc/viewPrd.asp?idproduct=50134
$97, is that expensive (especially considering the added value of the gpu)?

And your examples are a little skewed.
You should be comparing SoCs, not whole X86 systems with PPC SoCs.

But...I actually prefer a PPC solution myself.
The chips I'm interested in are low cost derivatives  Freescale's e5500 core, the T1022 and T1042.
At only a little over $100, the T1042 could give us a four core system at up to 1.4GHz that would fit very nicely in a board similar to the SAM460.
Those chips would also make awesome PPC accelerators.

[Iggy]

@ Kimmok,

I was psyched about the e5500 cored products when they were introduced, but these new chips have a really good price to performance ration.
And they should definitely have better performance than the AMCC cpus currently used by Acube.
I mentioned these to Trevor awhile ago, but they are a little below his target market.
They would be a perfect match for Acube's line though.
And we could use a good, low cost PPC board.

[Iggy]

So I guess that's what it is indeed, the contents of the OP, a rumor!!!

Ah, more Amiga vaporware.
How shocking! :hammer:

[Iggy]

They won't call it vaporware - they'll say they were only developing it for personal use and since a couple of boards were made then you can't call it vaporware

Prove "a couple of boards were made", and THEN it isn't BS.
And they didn't claim initially that they were developing it only for their personal use.
In fact, that would be rather pointless as the only soiftware that could easily be ported to it is either OS4 or MorphOS, neither of which is open source, so you would HAVE to make the boards available.

[Iggy]

There's photos of completed boards on their website.  Photoshop, some might claim, I suppose, but why go to all the trouble of assemble boards and then... what?  Not write the software to run them?  Is that what the holdup was?  Of course there's that whole issue of them having chosen a processor that's not compatible with any of the existing PPC software, which just made their job about a million times harder... *sigh*  I just don't understand people sometimes.

Well, I've been watching PPC processors for some time, and the processors that are compatible with the old boards are slowly being eol'd.
Add that to the fact that their is no documentation as to how most were built, well...
So, as I've said before, that leaves OS4 and MorphOS.
The former seems possible since the developers have been willing to support legacy hardware, but the latter is not that likely as many of the MorphOS developers don't even have Amiga hardware anymore.

And legacy PPC apps, are they really that important?
There aren't that many of them, and most can be run by an NG OS.

[Iggy]

I would pay good money for a PPC board that allows my A2000 to run OS4.1.

Yeah, I won't hold my breath.  :lol:

As an A2000 owner, I would not mind that myself, but then we'd also need a really good RTG card, as we don't have AGA graphics.

[Iggy]

Just put a PCI slot extension on the CPU card. Nice to dream......

PCI is a lot harder to interface to a 68K than most people realize.
The PT68K4 and PT68K5 boards I was using in the 80s and 90s had ISA slots which are much easier to connect to the 68K bus, but the cards available for that are more primitive.
I have two PCI bridge chips in my parts bins that are suited for use with 68K 040, 060, and earlier PPC cpus. They are still available, but they are quite complex and have some serious software requirements.

I spent a really long time considering building a system using those bridge chips, before desiding that the complexity of the needed design and the lack of decent 040 cpus made it too difficult to bother with considering the alternatives.

The last time I invested that much energy in something, I spent several months reworking a Freescale design to exchange the ALi Northbridge for a more common ATI SB600. While I was in the process of doing the board layouts, the MorphOS team released support for Mac PPC systems. As those ran just as fast or faster than what I was working on, it made continuing rather pointless.

Looking at what Jens has been talking about, makes me think we may soon be experiencing something similar.
What if instead of adding on to your A2000, you could just replace the board and have an 020/030 cpu and AGA graphics, with more memory, and maybe even a few PCI slots?