| From: | Sam Jordan |
| Date: | 14 Dec 99 at 19:28:48 |
| Subject: | Re: [warpup] Something about G3/G4 cards |
Am 14-Dez-99 schrieb Alin Ageu:
>Now why should we think (like Mr. Haage does) about licensing a "free"
>LinucPPC
>Box from IBM with a G3 proceessor. We could do better in supporting the P5
>concept. First of all i think that P5 will still deliver good work (they are
Supporting the P5 concept = turning away from AOS. And besides that, the
quality of these accelerator cards was far better in the past IMHO, but
I can not say whether this is their fault or whether the problem lies in
the AMIGA hardware itself, which experiences more and more problems with
such high-end components.
>So if the solution is using UAE for Amiga-Compatibility so why not using the
>hardware provided by a Amiga developer and profit of the increased
>capabilities
>of a G4 processor. I also believe that a QNX-version of UAE will be much
The processor is only as fast as the rest of the computer. The AMIGA
hardware architecture is old and outdated, support of peripheral devices
and other hardware is almost zero. For me it's no question that we
need to switch to another architecture.
>faster
>that a Linux-version.
That's utter *NONSENSE*. UAE is as fast as it is. Speed depends mainly
on the quality of the port, quality of the OS interfaces, optimizations,
compiler quality and runtime configuration of UAE.
>I also think that it is possible for P5 to make a standalone mainboard
instead
>of a accelerator card (something like the new Mac's with no serial/parallel
ABox .... PreBox ... #?Box ...
Making a motherboard is something very different than making an accelerator
card.
>The P5 hardware was never so uncompatible to other standards as always
>told(LinuxAPUS is the proof).
There was never such a horror design like most modern AMIGA's with all
those hardware components plugged in, permanently overheating or trashing
bits, far far away from the good old plug-and-play times (the preparation
at cologne was a *big* horror for me, just until we managed to make the
accelerator card and graphics card working). The AMIGA hardware architecture
simply can't handle such power anymore reliably IMHO. Additionally the whole
potential of the PowerPC is wasted for lots of applications due to lots of
bottlenecks, and this discrepancy will probably increase for even faster
processor. And as long as there is no AOSPPC we will never see the full
potential of the PPC hardware technology, except when looking to other
PPC-based architectures.
People should stop just focusing on processors, what we need is a good
hardware design as a whole, not only a processor. A processor for itself
is entirely useless, it *must* fit into an appropriate design.
I would definitely prefer a G3-based system based on a modern architecture
over a G4-based system based on a outdated/bottlenecked architecture.
>Lets stop complaining about all things and enter a common path for the future
>of the spirit of Amiga.(I think no programmer is willing to support 3
>standards
>if escena, P5,metabox and maybe H&P will do separate incompatible solutions
>for
>G3/G4 implementation)
H&P aims towards future OS development, which includes the move to AOSPPC,
and which also includes support for the POP hardware design, which represents
a modern design for modern hardware components/processors. And for once we
don't need to worry much about availability, as these boxes are made
primarly for the Linux community. We also won't have to worry about the
availability of hardware components for these computers, we only need to
worry about driver availability, but I guess that these should appear
as time goes on, even more as lots of the Linux drivers are becoming
Open Source in future and so can be used as template for writing AOS
drivers.
bye