Re: [Gomp-discuss] CIL representation ...

[Steven Bosscher]

Op wo 12-03-2003, om 11:14 schreef Lars Segerlund:
> 
>   I read the CIL doc's, and this sums it up.
>   CIL doesn't have everything needed to represent a parallell program at 
> the abstract level, it got locks and mutexes and non of the dependencies 
> needed for the variables. It's not a graph, it's a bytecode, so I think 
> we can drop it as a format interesting for gomp.

I do not share this opinion.  Maybe it doesn't have everything we need
(what exactly is missing do you think) but CIL is very generic, which
IMO is exactly what we need.

OTOH I haven't read Diego's Ph.D. thesis, maybe he used a less abstract
representation.

Either way, please be more specific.  It us *good* to write a 100 lines
long mail about *why* you think it "sums it up".  What are the things
you miss, which pieces do you like?

Oh, BTW, what's that line about dependencies?  Isn't it so that OpenMP
allows you to ignore any dependencies???  If not, you would need all 
the infrastructure of a auto-parallelizing compiler to validate the uses
of OpenMP directives...

Greetz
Steven