Implementing fine grain parallelism and automatic support for SIMD ISA extensions in gcc

[Martin Brain]

> Martin> I'm currently working for Bath University Computing Scienece
> Martin> department and we are interested in investigating fine grain parallelism
> Martin> in compiliation.  As far as I know no free compiler supports automatic
> Martin> synthesis of vector instructions, although icc does have support for
> Martin> this style of thing for P3 and P4 chips.  I would be trying to make this
> Martin> code as architecture independant as possible but aiming towards support
> Martin> for AltiVec, VIS, 3DNow and SSE first.
> Martin> Do people feel it is viable and worthwhile to attempt this?  Is anyone
> Martin> else working on this / thinking of doing so at the moment?  Would it be
> Martin> possible to integrate this with the existing support of vector data
> Martin> types?  Thanks for your thoughts.
> 	First step: read the GCC mailinglist as there already is a
> discussion underway.  It's the same as reading the existing journal
> literature before starting a project.
Sorry - I realised what I'd done just after sending it.  Sorry.

Cheers,
 - Martin