I am starting a new application(using C++, OOP/OOM)to be run on a Dual Quad Core server in production. Is it better touse CILK ++ or Intel's TBB for multicore support.
What are advantages of one over the other?
How will Intel integrate CILK ++ in their products?In which future product does CILK++ be included?
Actually, it's probably pretty hard to say without having some idea of the parallel algorithms you intend to use. Probably a more practical approach would be to try to prototype your code in each, some core element that you need the parallel boost (Amdahl's law and all that) and compare them. I haven't had a chance to test any Cilk++ code myself, so I really don't know how they compare. Regarding future plans, even if I did know anything, I probably couldn't tell you yet, but we're happy to have the Cilk Arts team on board.
