Prawn vs PDFKit Benchmark

benchmark_results.txt

-> % ./prawn_vs_pdfkit.rb 
                user     system      total        real
PDFKit:     1.760000   1.850000 231.550000 (286.579597)
Prawn:     99.140000   0.890000 100.030000 (102.124293)

2.4 GHz Intel Core 2 Duo
4 GB 1067 MHz DDR3

prawn_vs_pdfkit.rb

#!/usr/bin/env ruby
require 'rubygems'

require 'benchmark'
require 'faker'
require 'pdfkit'
require 'prawn'

LINES = (ENV["LINES"] ? ENV["LINES"].to_i : 300)
TIMES = (ENV["TIMES"] ? ENV["TIMES"].to_i : 100)
WIDTH = (ENV["WIDTH"] ? ENV["WIDTH"].to_i : 10)

LOREM = Faker::Lorem.paragraph(LINES)

def do_pdfkit
  doc = PDFKit.new(LOREM)
  doc.to_file('/tmp/prawn_vs_pdfkit/benchmark_pdfkit.pdf')
end

def do_prawn
  doc = Prawn::Document.new
  doc.text(LOREM)
  doc.render_file('/tmp/prawn_vs_pdfkit/benchmark_prawn.pdf')
end

Benchmark.bm(WIDTH) do |b|
  b.report("PDFKit:") { for i in 0..TIMES; do_pdfkit; end }
  b.report("Prawn: ") { for i in 0..TIMES; do_prawn; end }
end

You're right about not needing #freeze... I used for because according to this benchmark: http://www.ruby-doc.org/stdlib/libdoc/benchmark/rdoc/classes/Benchmark.html#M000010 it's a little faster, but times is cleaner, and the speed difference is probably trivial in this case. Thanks!

Your First Gist:

               user     system   total      real
PDFKit:    0.610000   0.770000  83.560000 (115.196074)
Prawn:     59.120000   0.340000  59.460000 ( 59.782119)

Your latest Gist:

            user        system    total       real
PDFKit:    0.590000   0.740000  82.560000 (111.891193)
Prawn:     57.630000   0.310000  57.940000 ( 57.975075)

Specs - Ruby 1.9.2p180
2.53GHz Intel Core i5
4 gb ram

Nice... looks like not doing extra superflous work improves performance time slightly :) (e.g. trying to freeze constants that hold fixnums...)

Seeing this kind of test, with our two machines, one generation apart was cool too. the i5 vs core 2 duo.

Cdog, i use pdfkit to print out an html invoice page. I couldn't figure out how to do it as easy as pdfkit... I'd love to see a test on that.