drake (version 5.0.0)

max_useful_jobs: Suggest an upper bound on the jobs in the next call to make(..., jobs = YOUR_CHOICE).

Description

The best number of jobs should be somewhere between 1 and max_useful_jobs(...). Any additional jobs more than max_useful_jobs(...) should be superfluous, and could even slow you down for make(..., parallelism = 'parLapply').

Usage

max_useful_jobs(config, imports = c("files", "all", "none"),
  from_scratch = FALSE)

Arguments

config

internal configuration list of make(...), produced also with drake_config().

imports

Set the imports argument to change your assumptions about how fast objects/files are imported. Possible values:

• 'all': Factor all imported files/objects into calculating the max useful number of jobs. Note: this is not appropriate for make(.., parallelism = 'Makefile') because imports are processed sequentially for the Makefile option.

• 'files': Factor all imported files into the calculation, but ignore all the other imports.

• 'none': Ignore all the imports and just focus on the max number of useful jobs for parallelizing targets.

from_scratch

logical, whether to assume the next make() will run from scratch so that all targets are attempted.

Value

A numeric scalar, the maximum number of useful jobs for make(..., jobs = ...).

Details

Set the imports argument to change your assumptions about how fast objects/files are imported. IMPORTANT: you must be in the root directory of your project.

See Also

vis_drake_graph, build_drake_graph, shell_file, drake_config()

Examples

# NOT RUN {
test_with_dir("Quarantine side effects.", {
load_basic_example() # Get the code with drake_example("basic").
config <- drake_config(my_plan) # Standard drake configuration list.
# Look at the graph. The work proceeds column by column
# in parallelizable stages. The maximum number of useful jobs
# is determined by the number and kind of targets/imports
# in the columns.
vis_drake_graph(config = config)
# Should be 8 because everythign is out of date.
max_useful_jobs(config = config) # 8
# Take into account targets and imported files.
max_useful_jobs(config = config, imports = 'files') # 8
# Include imported R objects too.
max_useful_jobs(config = config, imports = 'all') # 9
# Exclude all imported objects.
max_useful_jobs(config = config, imports = 'none') # 8
config <- make(my_plan) # Run the project, build the targets.
vis_drake_graph(config = config) # Everything is up to date.
# Ignore the targets already built.
max_useful_jobs(config = config) # 1
max_useful_jobs(config = config, imports = 'files') # 1
# Imports are never really skipped in make().
max_useful_jobs(config = config, imports = 'all') # 9
max_useful_jobs(config = config, imports = 'none') # 0
# Change a function so some targets are now out of date.
reg2 = function(d){
  d$x3 = d$x^3
  lm(y ~ x3, data = d)
}
vis_drake_graph(config = config)
# We have a different number for max useful jobs.
max_useful_jobs(config = config) # 4
# By default, max_useful_jobs() takes into account which
# targets are out of date. To assume you are building from scratch,
# consider using the "always" trigger.
max_useful_jobs(config, from_scratch = TRUE, imports = 'files') # 8
max_useful_jobs(config, from_scratch = TRUE, imports = 'all') # 9
max_useful_jobs(config, from_scratch = TRUE, imports = 'none') # 8
})
# }