Over at the Parallel for All blog, Mark Harris writes that Shared memory is a powerful feature for writing well optimized CUDA code. Access to shared memory is much faster than global memory access ...

NUMA (short for non-uniform memory access) is a computer memory architecture in which multiple processors are connected to the same shared memory pool via high-speed interconnections to control ...

system at the top of the market for high-end shared-memory computer. OpenMP shared-memory workloads will greatly benefit from this solution, as well as large in-memory computing and simulations.