REXML dom/pull/libxml benchmark plot with Ruby 3.3 YJIT

gistfile1.txt

# https://suer.hatenablog.com/entry/20110126/xml_parse の実験コードを参考にさせて頂きました。

require 'rubygems'
require 'gruff'
require 'benchmark'
require 'rexml/document'
require 'rexml/parsers/pullparser'
require 'xml/libxml'

# 最大子要素数
MAX = 10000
# 刻み
STEP = 100

g = Gruff::Line.new
RubyVM::YJIT.enable

def build_xml(n_elements, n_attributes)
  xml = '<?xml version="1.0"?><root>'
  n_elements.times do |i|
    xml << '<child '
    n_attributes.times {|j| xml << "id#{j}=\"#{i}\" " }
    xml << '/>'
  end
  xml << '</root>'
end

# XMLの大きさを順次変更し、yield で渡して解析時間を計測する
# 返り値は計測した時間の配列
def bench
  benchmarks = []
  0.step(MAX, STEP) do |i|
    xml = build_xml i, 2
    benchmarks << Benchmark.measure do
      yield xml
    end 
  end

  benchmarks.collect!{|b| b.total}
  puts benchmarks.last
  benchmarks
end

g.theme_37signals

puts "rexml(dom)"

# REXML で解析して plot 
g.data("rexml(dom)", 
  bench do |xml|
    REXML::Document.new(xml).elements.each("root/child") {|_|}
  end
)

puts "rexml(pull)"

g.data("rexml(pull)", 
  bench do |xml|
    parser = REXML::Parsers::PullParser.new(xml)
    while parser.has_next?
      parser.pull
    end
  end
)

puts "libxml-ruby(dom)"

# libxml-ruby で解析してplot
g.data("libxml-ruby(dom)", 
  bench do |xml| 
    XML::Document.string(xml).root.find('/root/child').each {|_|}
  end
)

l = {}
0.step(MAX, MAX / 4) do |i|
  l.merge!({i/STEP => i.to_s})
end
g.labels = l
g.x_axis_label = 'Number of lines'
g.y_axis_label = 'Seconds'
g.maximum_value = 0.2
g.minimum_value = 0
g.write('result.png')