An Experimentation Laboratory for the Automatic Parallelization of Programs written in the R Language (ALCHEMY)
Dr. Ing. Frank Padberg
- Person in Charge:Michael Mirold (Universität des Saarlandes)
The R language is being widely used infields such as bioinformatics and statistics, where it is common to process large data sets. R does not exploit the speed-up that may be gained from current multicore processors. In this thesis, I present the conception, the design, and the realization of the ALCHEMY parallelization laboratory and its integration into the R runtime environment.
I introduce the concept of Transmutators, small software components that apply transformations to an AST-based intermediate language called AIR, which is specically designed to facilitate analysis and code transformation.