Benchmark (Module)

In: benchmark.rb

The Benchmark module provides methods to measure and report the time used to execute Ruby code. Read on for illustrative examples.

Examples

Example 1

To measure the time to construct the string given by the expression "a"*1_000_000:

      require 'benchmark'

      puts Benchmark.measure { "a"*1_000_000 }

On my machine (FreeBSD 3.2 on P5100MHz) this reported as follows:

      1.166667   0.050000   1.216667 (  0.571355)

This report shows the user CPU time, system CPU time, the sum of the user and system CPU times, and the elapsed real time. The unit of time is seconds.

Example 2

To do some experiments sequentially, the bm method is useful:

      require 'benchmark'

      n = 50000
      Benchmark.bm do |x|
        x.report { for i in 1..n; a = "1"; end }
        x.report { n.times do   ; a = "1"; end }
        x.report { 1.upto(n) do ; a = "1"; end }
      end

The result:

             user     system      total        real
         1.033333   0.016667   1.016667 (  0.492106)
         1.483333   0.000000   1.483333 (  0.694605)
         1.516667   0.000000   1.516667 (  0.711077)

Example 3

Continuing the previous example, to put a label in each report:

      require 'benchmark'

      n = 50000
      Benchmark.bm(7) do |x|
        x.report("for:")   { for i in 1..n; a = "1"; end }
        x.report("times:") { n.times do   ; a = "1"; end }
        x.report("upto:")  { 1.upto(n) do ; a = "1"; end }
      end

The argument to bm (7) specifies the offset of each report according to the longest label.

The result:

                    user     system      total        real
       for:     1.050000   0.000000   1.050000 (  0.503462)
       times:   1.533333   0.016667   1.550000 (  0.735473)
       upto:    1.500000   0.016667   1.516667 (  0.711239)

Example 4

The times for some benchmarks depend on the order in which items are run. These differences are due to the cost of memory allocation and garbage collection.

To avoid these discrepancies, the bmbm method is provided. For example, to compare ways for sort an array of floats:

      require 'benchmark'

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

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

The result:

       Rehearsal -----------------------------------------
       sort!  11.928000   0.010000  11.938000 ( 12.756000)
       sort   13.048000   0.020000  13.068000 ( 13.857000)
       ------------------------------- total: 25.006000sec

                   user     system      total        real
       sort!  12.959000   0.010000  12.969000 ( 13.793000)
       sort   12.007000   0.000000  12.007000 ( 12.791000)

Example 5

To report statistics of sequential experiments with unique labels, benchmark is available:

      require 'benchmark'

      n = 50000
      Benchmark.benchmark(" "*7 + CAPTION, 7, FMTSTR, ">total:", ">avg:") do |x|
        tf = x.report("for:")   { for i in 1..n; a = "1"; end }
        tt = x.report("times:") { n.times do   ; a = "1"; end }
        tu = x.report("upto:")  { 1.upto(n) do ; a = "1"; end }
        [tf+tt+tu, (tf+tt+tu)/3]
      end

The result:

                    user     system      total        real
       for:     1.016667   0.016667   1.033333 (  0.485749)
       times:   1.450000   0.016667   1.466667 (  0.681367)
       upto:    1.533333   0.000000   1.533333 (  0.722166)
       >total:  4.000000   0.033333   4.033333 (  1.889282)
       >avg:    1.333333   0.011111   1.344444 (  0.629761)

Methods

benchmark   bm   bmbm   measure   realtime  

Constants

BENCHMARK_VERSION = "2002-04-25"
  BENCHMARK_VERSION is version string containing the last modification date (YYYY-MM-DD).
CAPTION = Benchmark::Tms::CAPTION
  The default caption string (heading above the output times).
FMTSTR = Benchmark::Tms::FMTSTR
  The default format string used to display times. See also Benchmark::Tms#format.

Classes and Modules

Class Benchmark::Job
Class Benchmark::Report
Class Benchmark::Tms

Public Instance methods

Reports the time required to execute one or more blocks of code.

Note: Other methods provide a simpler interface to this one, and are suitable for nearly all benchmarking requirements. See the examples in Benchmark, and the bm and bmbm methods.

Example:

    require 'benchmark'
    include Benchmark          # we need the CAPTION and FMTSTR constants

    n = 50000
    Benchmark.benchmark(" "*7 + CAPTION, 7, FMTSTR, ">total:", ">avg:") do |x|
      tf = x.report("for:")   { for i in 1..n; a = "1"; end }
      tt = x.report("times:") { n.times do   ; a = "1"; end }
      tu = x.report("upto:")  { 1.upto(n) do ; a = "1"; end }
      [tf+tt+tu, (tf+tt+tu)/3]
    end

The result:

                    user     system      total        real
       for:     1.016667   0.016667   1.033333 (  0.485749)
       times:   1.450000   0.016667   1.466667 (  0.681367)
       upto:    1.533333   0.000000   1.533333 (  0.722166)
       >total:  4.000000   0.033333   4.033333 (  1.889282)
       >avg:    1.333333   0.011111   1.344444 (  0.629761)

The parameters accepted are as follows:

caption:A string printed once before execution of the given block.
label_width:An integer used as an offset in each report.
fmtstr:A string used to format each measurement. See Benchmark::Tms#format.
labels:The remaining parameters are used as prefix of the format to the value of block; see the example above.

This method yields a Benchmark::Report object.

[Source]

# File benchmark.rb, line 198
  def benchmark(caption = "", label_width = nil, fmtstr = nil, *labels) # :yield: report

    sync = STDOUT.sync
    STDOUT.sync = true
    label_width ||= 0
    fmtstr ||= FMTSTR
    raise ArgumentError, "no block" unless iterator?
    print caption
    results = yield(Report.new(label_width, fmtstr))
    Array === results and results.grep(Tms).each {|t|
      print((labels.shift || t.label || "").ljust(label_width), 
            t.format(fmtstr))
    }
    STDOUT.sync = sync
  end

A simple interface to benchmark, bm is suitable for sequential reports with labels. For example:

    require 'benchmark'

    n = 50000
    Benchmark.bm(7) do |x|
      x.report("for:")   { for i in 1..n; a = "1"; end }
      x.report("times:") { n.times do   ; a = "1"; end }
      x.report("upto:")  { 1.upto(n) do ; a = "1"; end }
    end

The argument to bm (7) specifies the offset of each report according to the longest label.

This reports as follows:

                    user     system      total        real
       for:     1.050000   0.000000   1.050000 (  0.503462)
       times:   1.533333   0.016667   1.550000 (  0.735473)
       upto:    1.500000   0.016667   1.516667 (  0.711239)

The labels are optional.

[Source]

# File benchmark.rb, line 239
  def bm(label_width = 0, *labels, &blk) # :yield: report

    benchmark(" "*label_width + CAPTION, label_width, FMTSTR, *labels, &blk)
  end

Similar to bm, but designed to prevent memory allocation and garbage collection from influencing the result. It works like this:

  1. The rehearsal step runs all items in the job list to allocate enough memory.
  2. Before each measurement, invokes GC.start to prevent the influence of previous job.

If the specified label_width is less than the width of the widest label passed as an argument to item, the latter is used. (Because bmbm is a 2-pass procedure, this is possible.) Therefore you do not really need to specify a label width.

For example:

      require 'benchmark'

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

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

The result:

       Rehearsal -----------------------------------------
       sort!  11.928000   0.010000  11.938000 ( 12.756000)
       sort   13.048000   0.020000  13.068000 ( 13.857000)
       ------------------------------- total: 25.006000sec

                   user     system      total        real
       sort!  12.959000   0.010000  12.969000 ( 13.793000)
       sort   12.007000   0.000000  12.007000 ( 12.791000)

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

[Source]

# File benchmark.rb, line 283
  def bmbm(width = 0, &blk) # :yield: job

    job = Job.new(width)
    yield(job)
    width = job.width
    sync = STDOUT.sync
    STDOUT.sync = true

    # rehearsal

    print "Rehearsal "
    puts '-'*(width+CAPTION.length - "Rehearsal ".length)
    list = []
    job.list.each{|label,item|
      print(label.ljust(width))
      res = Benchmark::measure(&item)
      print res.format()
      list.push res
    }
    sum = Tms.new; list.each{|i| sum += i}
    ets = sum.format("total: %tsec")
    printf("%s %s\n\n",
           "-"*(width+CAPTION.length-ets.length-1), ets)
    
    # take

    print ' '*width, CAPTION
    list = []
    ary = []
    job.list.each{|label,item|
      GC::start
      print label.ljust(width)
      res = Benchmark::measure(&item)
      print res.format()
      ary.push res
      list.push [label, res]
    }

    STDOUT.sync = sync
    ary
  end

Returns the time used to execute the given block as a Benchmark::Tms object.

[Source]

# File benchmark.rb, line 326
  def measure(label = "") # :yield:

    t0, r0 = Benchmark.times, Time.now
    yield
    t1, r1 = Benchmark.times, Time.now
    Benchmark::Tms.new(t1.utime  - t0.utime, 
                       t1.stime  - t0.stime, 
                       t1.cutime - t0.cutime, 
                       t1.cstime - t0.cstime, 
                       r1.to_f - r0.to_f,
                       label)
  end

Returns the elapsed real time used to execute the given block.

[Source]

# File benchmark.rb, line 341
  def realtime(&blk) # :yield:

    Benchmark::measure(&blk).real
  end

[Validate]