Pavan Balaji a e'crit :
I'm mostly concerned whether we will have enough resources and/or experience in the working group to look into the other models at all. Whil it's a good goal, it might not be realistic. If we don't, we might need to restrict the goals of the working group to something smaller. It's a good point. I think it will largely depend on the experience of people participating in the WG. As Pavan suggested during the Telecon previously, we can send emails about the interfaces/paradigm we want to specifically investigate. As for our group in INRIA, we've been working for quite some time on runtime systems tailored for multithreading. So we are willing to look at this. We also have some people working on the design of a kind of CUDA task scheduler so we are also defintively interested by this point too.
Guillaume