Due to high computational demands, the inference of large phylogenetic trees from molecular sequence data requires the use of HPC systems in order to obtain the necessary computational power and memory. The continuous explosive accumulation of molecular data, which is driven by the development of cost-effective sequencing techniques, amplifies this requirement additionally. Furthermore, a continuously increasing degree of parallelism is necessary in order to exploit the performance of emerging multi-core processors efficiently.
This dissertation describes scalable parallelization schemes for the inference of large phylogenetic trees as well as tangible implementations of those which also eliminate memory requirements as a limiting factor for phylogenetic analyses. Additionally, it pinpoints the properties of current multi-core shared and distributed memory architectures and describes novel approaches for their efficient exploitation.
«
Due to high computational demands, the inference of large phylogenetic trees from molecular sequence data requires the use of HPC systems in order to obtain the necessary computational power and memory. The continuous explosive accumulation of molecular data, which is driven by the development of cost-effective sequencing techniques, amplifies this requirement additionally. Furthermore, a continuously increasing degree of parallelism is necessary in order to exploit the performance of emerging m...
»