Rth: Rth: Parallel R through Thrust

#' Parallel Computation of Histograms
#'
#' Similar to R's \code{hist()}
#'
#' Similar to \code{hist()}, though with fewer options.
#'
#' @param xname
#' Name of data vector, quoted.
#' @param nbins
#' The number of bins.
#' @param plot
#' If TRUE, plot will be displayed.
#' @param nthreads
#' An optional argument for specifying the number of threads
#' (in the case of OpenMP or TBB backends) that the routine should (at most)
#' use.  Does nothing with a CUDA backend.  See \link{nthreads}.
#' @param nchunks
#' Number of chunks for partitioning the data.
#'
#' @return Object of R class \code{"histogram"}.
#'
#' @examples
#'
#' \dontrun{
#' # R built-in geyser data
#' fw <- faithful$waiting
#' fwh <- rthhist("fw")
#' fwh  # display plot, famously bimodal
#' fwh <- rthhist("fw",plot=FALSE)
#' fwh
#' # bin counts, left endpt, binwidth printed out
#' }
#'
#' @export
# rthhist <- function(xname,nbins=15,plot=TRUE, nthreads=rth.nthreads(),nchunks=nthreads) {
#
#   # in TBB cases, automatic() returns -1, the code for giving TBB carte
#   # blanche
#   if (nthreads == -1) nchunks <- 2
#
#   if (mode(xname) != "character")
#      stop("the data vector must be specified by quoted name")
#   x <- get(xname)
#   if (!is.double(x))
#     storage.mode <- "double"
#
#   tmp <- .Call(c_rthhist, x,
#     as.integer(nbins), as.integer(nchunks), as.integer(nthreads))
#
#   if (!plot)
#     return(list(counts=tmp$counts, left=tmp$left, binwidth=tmp$binwidth))
#
#   plt <- list()
#   class(plt) <- "histogram"
#   plt$breaks <- seq(from=tmp$left,by=tmp$binwidth,length.out=nbins+1)
#   plt$counts <- tmp$counts
#   plt$xname <- xname
#   plot(plt)
# }

NULL