Benchmark the package — benchmark • predped

Function that allows users to benchmark predped on their own device. Times are measured through repeated assessments with Sys.time(), providing us with a measure of the user-based time for the functions. Importantly, this is different from pure execution time of the function, which may be less than the user time.

Usage

benchmark(
  x = NULL,
  iterations = 100,
  summarize = TRUE,
  digits = 2,
  progress = TRUE
)

Arguments

x: Character or character vector containing the names of the functions to include in the benchmark. Functions should be part of the package. Defaults to NULL, denoting that all functions will be tested.
iterations: Integer denoting the number of times to repeat each function's benchmark case. The higher this number, the longer the benchmark will take but also the more accurate the distribution of times will be per function. Defaults to 100.
summarize: Logical denoting whether to summarize the results of the benchmark. If TRUE, the function will output an HTML file containing the summarized results of the benchmark (mean, standard deviation, and 2.5 will be a list of lists containing the raw execution times per function and per benchmark case for the function. Defaults to TRUE.
digits: Integer denoting the decimal points to which to round to. Ignored if summarize = FALSE. Defaults to 2.
progress: Logical denoting whether to add a progress bar for the benchmarks. Defaults to TRUE.

Value

Either a list of lists containing the raw execution times per benchmark case per function (if summarize = FALSE) or an HTML file summarizing the results of the benchmark (if summarize = TRUE)

Examples

if (FALSE) { # \dontrun{
# Run benchmark for in_object with 100 iterations
benchmark(
    x = "in_object",
    iterations = 100
)

# Run benchmark for all functions with 10 iterations
benchmark(
   iterations = 10
)
} # }