Over the past decade, data mining techniques has been employed on a wide range of shared and distributed memory architectures, using a variety of parallel and distributed implementations over the past decade. Originally motivated by the large size and distributed nature of the data, the success of these approaches is, in part, founded upon the fairly homogeneous nature of the underlying architectures. Recently, from a high-performance computing viewpoint, there has been a shift towards the use of heterogeneous architectures. This shift is evident, for example, in the development of the Cell Broadband Engine Architecture and its use in RoadRunner, and the utilization of GPUs as accelerators in scientific computing. We will review such architectures and then discuss their applicability in High-Performance Data Mining, using recent work on the Nearest Neighbour algorithm, Decision Trees, and Self-Organizing Maps. We will also briefly discuss a current data-mining problem in astronomy, as well as related simulations.