vectorizeFunction: Create a vectorized version of a scalar function

In duncantl/RLLVMCompile: LLVM-based compiler for R code

Description

This function takes an R scalar function or a compiled version of an R function and generates a new function that loops over its inputs element at a time and calls the original scalar function. In the near future, this function will generate a compiled version of the vectorized version. This will be merely a compilation of a call to sapply.

Usage

vectorizeFunction(f, fc = NULL, typeInfo = NULL, module = as(fc, "Module"),
                  scalarFunName = getName(fc), vectorArgName = names(parms)[1])

Arguments

f	the scalar R function
fc	the compiled version of this scalar function
typeInfo	information about the types of the scalar routine/function
module	the LLVM module in which to create the new routine
scalarFunName	the name of the compiled scalar routine
vectorArgName	the name of the parameter over which we want to vectorize the computations

Value

The compiled routine

Author(s)

Duncan Temple Lang

See Also

compileFunction

Examples

Dnorm = function(x, mu = 0, sd = 1)
         1/sqrt(2 * pi * sd^2) * exp( - .5*( (x-mu)/sd ) ^2)
fc = compileFunction(Dnorm, DoubleType, list(DoubleType, DoubleType,
         DoubleType))

g = vectorizeFunction(Dnorm, fc) 

duncantl/RLLVMCompile documentation built on May 15, 2019, 5:31 p.m.