The fbseqCUDA package is an internal backend of fbseq package that runs the Markov chain Monte Carlo (MCMC) procedure behind the scenes. It is implemented with CUDA for acceleration with parallel computing. For installation, CUDA must be installed. To use fbseqCUDA package in an MCMC, fbseq package must be installed, and a CUDA-capable general-purpose graphics processing unit (GPU) must be installed on your machine.
- The R version and R packages listed in the "Depends", "Imports", and "Suggests" fields of the "package's DESCRIPTION file.
- A CUDA-capable NVIDIA graphics processing unit (GPU) with compute capability 2.0 or greater.
- CUDA version 6.0 or greater. More information about CUDA is available through NVIDIA.
- Optional: the code uses double precision values for computation, so GPUs that natively support double precision will be much faster than ones that do not.
Navigate to a list of stable releases on the project's GitHub page. Download the desired tar.gz bundle, then install it either with install.packages(..., repos = NULL, type="source") from within R R CMD INSTALL from the Unix/Linux command line.
For this option, you need the devtools package, available from CRAN or GitHub. Open R and run
library(devtools)
install_github("wlandau/fbseqCUDA")
Open a command line program such as Terminal in Mac/Linux and enter the following commands.
git clone [email protected]:wlandau/fbseqCUDA.git
R CMD build fbseqCUDA
R CMD INSTALL ...
where ... is replaced by the name of the tarball produced by R CMD build.
If CUDA is not found, open fbseqCUDA/src/Makevars in a text editor. The top line reads
CUDA_HOME = /usr/local/cuda
but this may not be correct for your system. Replace /usr/local/cuda with the correct path to the installation of CUDA on your computer.