Multicore Tools

If you want the performance advantages of multicore processing, you have to make some choices. For instance, do you invest in optimizing for the Cell, or target a GPU from NVIDIA? Ideally, you don't want to have to make that choice. Jon Erickson recently talked with Michael McCool, chief scientist and founder of RapidMind. RapidMind provides a platform that works with existing C++ compilers and employs the concept of data parallelism to exploit the benefits of multiple multicore implementations. According to McCool, this hides the complexity of optimization behind a simple API. Check out the interview here.

Posted by Kevin Carlson at 01:42 PM  Permalink