SPRINT Project
2014-Jan-30 15:22 UTC
[R-pkgs] SPRINT 1.0.5 release (parallelised R functions for complex analysis and large data sets)
Dear All We have recently released SPRINT v1.0.5. SPRINT provides parallelised versions of existing R functions (e.g. statistical, machine learning, utility) that often exceed computational limits (speed or memory) with large or complex data sets. SPRINT 1.0.5 now: -> runs on Mac OSX multi-cores (i.e. you may not need to obtain access to compute clusters) -> provides a parallelised version of the Hamming distance (based on original version contained in stringdistmatrix() by M van der Loo, and parallelised differently from the option provided therein) -> compliant with MPI3 (and the latest version of the mpich package) Download and instructions available at www.r-sprint.org. SPRINT 1.0.5 is not available on CRAN as we have not yet completed testing against the latest major release of R (3.0.x). However, we will release a small incremental update of SPRINT as version 1.0.6 within the next few weeks that will be available via CRAN and will work with R3. Notes: - With the availability of SPRINT for Mac users, we hope to address computational issues that are too large or too slow for these users, but still small enough to not require a much larger number of nodes and cores (i.e. compute clusters or supercomputers like HECToR or ARCHER). - We have parallelised the Hamming distance as from collaborators and course participants we know this to have potential in use for Next-Gen-Sequencing, e.g. measuring pairwise distances between nucleotide sequences. In principle, though, Hamming is a distance useful for string or binary data vectors. SPRINT Project sprint@ed.ac.uk www.r-sprint.org The University of Edinburgh -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. _______________________________________________ R-packages mailing list R-packages@r-project.org https://stat.ethz.ch/mailman/listinfo/r-packages