Supercomputing for Me = Personal Supercomputing
With available many-core processing technologies, e.g. GPGPU and x86-based platforms, substantial amounts of workloads can be nowadays processed on so-called Personal Supercomputer systems which are attractive in particular for SME business.
Under the umbrella of the SC4Me project we investigate different algorithms for small configurations of advanced many-core platforms.
Similarity Search in High-Dimensional Spaces
The project aims to implement an advanced algorithm for kNN search in high-dimensional space, e.g. large time-series data sets, on GPGPUs.
(Patrick Schäfer, Thomas Steinke )
In collaboration with Ch. Gillan from Queen's University Belfast we investigated the performance of a numerical quadrature algorithm in atomic physics code on GPGPU and FPGA accelerator platforms.