This is joint work with Dr. Michael Monagan at Simon Fraser University. We present our recent work of scaling Maple's polynomial algorithms to multicore CPUs, including parallel algorithms with superlinear speedup and the new polynomial data structure that we have added to the kernel. The larger issue we hope to discuss is how to scale different computer algebra systems to future processors. Our strategy for Maple reflects its design as a small, compact kernel with high level library routines that scales down as well as up. The ideal strategy for Magma could be quite different, however I don't presume to know.