makeParallel: Make Parallel Code From Serial

In clarkfitzg/codedoctor: Transform Serial R Code into Parallel R Code

Description

makeParallel is a high level function that performs all the steps to generate parallel code.

Usage

makeParallel(code, isFile = file.exists(code), expr = if (isFile)
  parse(code, keep.source = TRUE) else parse(text = code, keep.source =
  FALSE), data = NULL, nWorkers = parallel::detectCores(),
  platform = Platform(nWorkers = nWorkers), run = FALSE,
  scheduler = schedule, ..., generator = generate,
  generatorArgs = list(), outFile = FALSE, prefix = "gen_",
  overWrite = FALSE)

Arguments

code	file name or a string containing code to be parsed
isFile	logical, is the code a file name?
expr	expression, for example from parse
data	list of data descriptions. Each element is a DataSource. The names of the list elements correspond to the variables in the code that these objects are bound to.
nWorkers	integer, number of parallel workers
platform	Platform describing resource to compute on
run	logical, evaluate the code once to gather timings?
scheduler,	function to produce a Schedule from a TaskGraph.
...	additional arguments to schedule methods
generator	function to produce GeneratedCode from a Schedule
generatorArgs	list of named arguments to use with generator
outFile	character name of the file to write the generated script. If FALSE then don't write anything to disk. If TRUE and code comes from a file then use prefix to make a new name and write a script.
prefix	character added to front of file name
overWrite	logical write over existing generated file

Details

The following are the high level steps:

1. Infer the task graph

2. Schedule the statements

3. Generate parallel code

The arguments allow the user to control every aspect of this process. For more details see vignette("makeParallel-concepts").

Value

code object of class GeneratedCode

Examples

# Make an existing R script parallel
script <- system.file("examples/mp_example.R", package = "makeParallel")
makeParallel(script)

# Write generated code to a new file
newfile <- tempfile()
makeParallel(script, file = newfile)

# Clean up
unlink(newfile)

# Pass in code directly
d <- makeParallel(expr = parse(text = "lapply(mtcars, mean)"))

# Examine generated code
writeCode(d)

# Specify a different scheduler
pcode <- makeParallel("x <- 1:100
y <- rep(1, 100)
z <- x + y", scheduler = scheduleTaskList)

# Some schedules have plotting methods
plot(schedule(pcode))

clarkfitzg/codedoctor documentation built on Nov. 18, 2020, 4:34 p.m.