bmbm

Sometimes benchmark results are skewed because code executed earlier encounters different garbage collection overheads than that run later. #bmbm attempts to minimize this effect by running the tests twice, the first time as a rehearsal in order to get the runtime environment stable, the second time for real. GC.start is executed before the start of each of the real timings; the cost of this is not included in the timings. In reality, though, there’s only so much that #bmbm can do, and the results are not guaranteed to be isolated from garbage collection and other effects.

Because #bmbm takes two passes through the tests, it can calculate the required label width.

require 'benchmark'

array = (1..1000000).map { rand }

Benchmark.bmbm do |x|
  x.report("sort!") { array.dup.sort! }
  x.report("sort")  { array.dup.sort  }
end

Generates:

Rehearsal -----------------------------------------
sort!   1.440000   0.010000   1.450000 (  1.446833)
sort    1.440000   0.000000   1.440000 (  1.448257)
-------------------------------- total: 2.890000sec

            user     system      total        real
sort!   1.460000   0.000000   1.460000 (  1.458065)
sort    1.450000   0.000000   1.450000 (  1.455963)

#bmbm yields a Benchmark::Job object and returns an array of Benchmark::Tms objects.