This is just an internal utility included in the package which is not designed to be accurate. If you need accurate benchmarks, you should take a look at the microbenchmark and bench R packages.

benchmarkito(..., times = 100, rand_ord = TRUE)

Arguments

...

List of expressions to benchmark.

times

Integer scalar. Number of replicates.

rand_ord

Logical. When TRUE, the expressions are executed in a random order.

Value

A data frame of class ergmito_benchmark with times rows and the following columns:

  • id Integer. Id of the expression.

  • expr Factor. Expression executed.

  • user.self, sys.self, elapsed, sys.child time in seconds (see proc.time()).

Includes the following attributes: ncalls, call, label, and expr.

Details

The print method prints a summary including quantiles, relative elapsed times, and the order (fastest to slowest).

Examples

bm <- benchmarkito( exp(1:100000), sqrt(1:100000), times = 20 ) plot(bm)
print(bm)
#> call: #> benchmarkito(exp(1:1e+05), sqrt(1:1e+05), times = 20) #> #> Times (in seconds, see proc.time()): #> #> label user.self sys.self elapsed user.child sys.child 5% 50% #> 1 exp(1:1e+05) 0.00105 7e-04 0.00170 0 0 0.001 0.002 #> 2 sqrt(1:1e+05) 0.00075 3e-04 0.00105 0 0 0.000 0.001 #> 95% relative ord #> 1 0.00200 1.619048 2 #> 2 0.00125 1.000000 1