power pc

[forsyth@plan9.cs.york.ac.uk forsyth@plan9.cs.york.ac.uk]

>>And, for a technical question, is anyone planning a PowerPC port?

yes.  to some extent, it isn't really a technical question though, but political and financial.
the technical part is relatively straightforward, given the right information.
the political aspect is that it is difficult to obtain any information
about the RS/6000 machines i've got to hand.

i looked at the IBM Microelectronics PowerPC Design Kit, but even though
it (apparently) comes with reasonably precise hardware documentation, it
was inferior in several ways to the Indy I finally bought.
once I took into consideration all the extra bits required, it was
fairly expensive (at least in Britain).  the financial aspect is
that i could afford only to buy the Indy.  nevertheless, I have done
a working subset of the assembler, linker and compiler for the RS6000 processor,
and intend to complete them.  although i was quite enthusiastic about the Power architecture to start with,
the syntax of the PowerPC's order code (at the binary level) is unpleasant.
give me a MIPS or an Alpha any day.

still, if IBM ever gets itself together (or the clone manufacturers do),
the Power PC might offer a cheap but fast platform, without the crud
that goes with the Intel stuff.  (best of all, you get a few registers
you can actually use.)  despite the comments above about obtaining
information, they also do seem to be publishing more details
about the architecture as books, independently of the Design Kit.
unfortunately, architectural details do not apply precisely to a given
implementation.  (think of Sun and their wretched no-two-the-same-colour Sparcs.)

although i'm happy to complete the compiler/assembler/linker, having started,
since that's easy to do using the Power 250 in the department, you might
have to do the kernel ports yourself!