|Re: CISC to RISC translators ? Roger@natron.demon.co.uk (1994-08-23)|
|From:||Roger@natron.demon.co.uk (Roger Barnett)|
|Keywords:||architecture, translator, UNCOL|
|Date:||Tue, 23 Aug 1994 07:48:07 GMT|
roedy@BIX.com "Roedy Green" writes:
> Difficulties moving from CISC to RISC:
> [list snipped]
There is also the question of licencing of the software being ported - if
someone used Digital's VEST to move one of our products from the VAX to an
Alpha without asking us first then we would not be impressed !
On the subject of VEST, I've heard that Digital are expandings its range
to handle migrations (to the Alpha only, I should think) from non-Digital
platforms such as Sparc - anyone got anything definite on this ?
> A more promising approach is to use a P-CODE, designed to be easy to
> translate both into native CISC or native RISC architectures. It would be
> like a super CISC instruction set, with extra baggage to aid in the
> translation. It could then be translated/compiled as needed at load time.
> This would give the advantage of compact, fast loading executables that
> could run on many platforms. By making the P-CODE set very CISCy it gives
> plenty of room for RISC to optimise WITHIN each instruction.
This approach sounds like the Transportable Distribution Format from
the DRA at Malvern in the UK (part of which was going to be adopted by the
OSF for their Architecture Neutral Distribution Format - whatever happened
to that ?). This provides for a single set of sources, with no conditional
compilations, to be compiled once into an intermediate form and then run
through an "installer" on each target system to create a native executable.
The last time I spoke to the people at DRA these tools were only available
to developers rather than users, and were very expensive (although I believe
there are other groups in Europe also working on TDF).
[We've had discussions of TDF and ANDF in the past. I gather there have
been some promising demos, but it's not a product yet. And there's the
uncomfortable fact that every previous attempt to build a universal
intermediate language, back to the original UNCOL project in the 1950s,
Return to the
Search the comp.compilers archives again.