Boundaries of Supercomputing

Abstract

Supercomputers can reach an unprecedented degree of scale and miniaturization reaches the quantum level of manufacturing. Non-regarding this progress, however, computing capabilities are and will remain insufficient to meet the demands of many compute intensive scenarios. The major obstacle thereby consists in the non-deterministic polynomial (NP) nature of these problems. Recent research and development seems to have almost forgotten about this intrinsic problem. With this paper we want to remind of the relevance of NP for supercomputing by exploring its impact on future computing development and discuss potential approaches to relieving (not solving) this issue.