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

Re: Intel: timely article



Hi,


A.E.Lawrence wrote:
> > There are papers here:
> > http://www.cs.kuleuven.ac.be/conference/EuroSys2006/papers/p177-fahndrich.pdf
>
> Following myself up, Fred's RMoX is in the paper: see p35 . And they
> complain that it isn't completely written in occam! Agreed :-)


I'd agree with some of that too ;).  I've toyed around a bit with generating
the run-time kernel from occam-pi sources, should be possible, but might
need special (safe) magic in the language to keep some of the existing
optimisations.  An exercise for the new compiler at some point. :)

Incidently, that paper makes an incorrect claim about occam regarding
ALTs:

    "Furthermore, an OCCAM thread waits on a single channel, whereas
     our threads can be blocked on any number of wait conditions, which
     causes more accounting on waking a thread."

I pointed that out to the authors at the conference, hopefully to be
filtered into to future publications.


Cheers,

-- Fred

Attachment: pgpWuyrowZs25.pgp
Description: PGP signature