[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Universal Processor (was Re: [oc] x86 IP Core)

Extending my original post a bit further:

A compiler to generate a core which translates
an arbitrary instruction set to another arbitrary
instruction set.  Optionally, each instruction set
may be specified using the 'gcc machine description'
language, or similar.  One advantage of using the gcc format is
that the machine description can be fed into gcc,
so each instruction set comes with a 'free' compiler
and one is guaranteed the instruction set matches the
compiler, since they came from the same source.

As a side note, perhaps such an architecture would
not infringe on 'rights' of other companies?  As far
as I know, it is not illegal to describe an instruction set.
The 'translator compiler' would be a novel stand alone
development (possibly worthy of its own patent).  Such
a compiler could hardly infringe a patent on the
implementation of a microprocessor instruction, as
its behaviour is entirely dependent on any machine
description fed to it.  Claiming infringement would
be like claiming a VHDL compiler infringes since
given the right source file as input it can produce
an 'illegal' microprocessor.

To tell you the truth (and as pointed out by others), I
suspect this would be a very difficult project, possibly
not even practical.  I've posted this just so the idea is
entered into the mailing list archive and can serve as
prior art if it hasn't already been patented.

(Maybe we should have a 'crazy ideas' mailing list?  Does
anyone know whether such a mailing list archive would hold
up as prior art in court?  I guess the 'wishlist' is along
these lines, but it is not very interactive.  Maybe a
'wiki' wishlist would be better?)


Incidentally, the existing wishlist really needs a clean out.
I'm willing to give it a go sometime (probably during
October given my current 'to do' list), if someone can give
me the necessary permissions (username: jdalton).  I would
start by archiving the old wishlist, so we don't lose anything,
then pruning all the 'crud' out of it.  Is there a database
behind the wishlist page, or is the page itself the database?
To unsubscribe from cores mailing list please visit http://www.opencores.org/mailinglists.shtml