Mark is a researcher in the Scientific Data Group, in the Computer Science and Mathematics Divisionof Oak Ridge National Laboratory. His work at Oak Ridge is focused on in situ analysis and visualization. In the high performance computing world, there is an increasingly large disconnect between multiple levels of a supercomputer (compute vs memory vs disk vs network storage). This has a knock-on effect for scientific visualization: previously, computational simulations were stored to a network file system and visualization of the data was performed on a separate visualization cluster in a post hoc manner. The bandwidth problem has forced a rethink of how scientific visualization is performed where the visualization needs to be coupled to the simulation. With the advent of heterogenous supercomputers, such as OLCF Titan which uses GPU accelerators, scientific visualization codes need to adaptability to this ever changing landscape. One way to address this issue is with toolkits that are write-once/run-anywhere: in other words, VTK-m.