iFes: An incremental feature selection algorithm for gene expression profile accelerated by NVIDIA GPU.(some install hints are referred to R package gputools.)

=========
Prerequisites for iFes:

1. A CUDA capability 1.3 or higher NVIDIA GPU, you can find GPU list which support CUDA program framework.
2. NVIDIA CUDA 5.5 or higher toolkit, you can get from https://developer.nvidia.com/cuda-downloads.

To install the package, you must first install Nvidia's CUDA 5.5 toolkit available from

=========

Installation Hints

Set and export the environment variable CUDA_HOME. In most cases, something like this will do:

export CUDA_HOME=/usr/local/cuda/5.5

Set your LD_LIBRARY_PATH to include CUDA's library directory. In most cases, something like this will do:

export LD_LIBRARY_PATH=/usr/local/cuda/5.5/lib64:$LD_LIBRARY_PATH

Use R to install the source package. For example, something like this will do:

R CMD INSTALL iFes*.tar.gz

You can edit the file iFes/src/config.mk to suit your particular environment if the installation failed.


