- author = "William B. Langdon",
- title = "Large Scale Bioinformatics Data Mining with Parallel Genetic Programming on Graphics Processing Units",
- booktitle = "Massively Parallel Evolutionary Computation on {GPGPUs}",
- publisher = "Springer",
- year = "2013",
- editor = "Shigeyoshi Tsutsui and Pierre Collet",
- series = "Natural Computing Series",
- chapter = "15",
- pages = "311--347",
- keywords = "genetic algorithms, genetic programming, GPU",
- isbn13 = "978-3-642-37958-1",
-
URL = "
http://www.springer.com/computer/ai/book/978-3-642-37958-1",
-
DOI = "
doi:10.1007/978-3-642-37959-8_15",
- size = "36 pages",
- abstract = "The NCBI GEO GSE3494 breast cancer dataset contains hundreds of Affymetrix HG-U133A and HG-U133B GeneChip biopsies each with a million variables. Multiple genetic programming (GP) runs on a graphics processing unit (GPU) hardware, each with a population of five million programs both winnow (select) useful variables from the chaff and evolve small (three inputs) data models. The SPMD CUDA interpreter exploits the GPU's single instruction multiple data SIMD mode of parallel computing, even though the GP populations contain different programs. A 448 node nVidia Fermi C2050 Tesla graphics card delivers 8.5 giga GPops per second. In addition to describing our implementation, we survey current GPGPU work in bioinformatics and genetic programming.",
-
notes = "
Table 3 pages 338-339, gives published speeds in terms of millions of GPops per second of genetic programming on GPUs.",
