Sha256: 56a84f0f30fcbffa1e5ccb8eb33b5f980c7bc8913372220fefeb2cf74009c0fa
Contents?: true
Size: 1.88 KB
Versions: 18
Compression:
Stored size: 1.88 KB
Contents
require 'benchmark/ips'
require 'stringio'
puts "Ruby #{RUBY_VERSION} at #{Time.now}"
puts
Benchmark.ips do |x|
# These two tests look at the cost of simply writing to a buffer.
x.report("String\#<<") do |i|
buffer = String.new
while (i -= 1) > 0
buffer << "String #{i}"
end
end
x.report("Array\#<<") do |i|
buffer = []
while (i -= 1) > 0
buffer << "String #{i}"
end
buffer.join
end
x.report("StringIO") do |i|
buffer = StringIO.new
while (i -= 1) > 0
buffer << "String #{i}"
end
end
x.compare!
end
# Calculating -------------------------------------
# String (Amortized) 91666 i/100ms
# StringIO (Amortized) 69531 i/100ms
# -------------------------------------------------
# String (Amortized) 2856024.9 (±8.8%) i/s - 14208230 in 5.017309s
# StringIO (Amortized) 2424863.3 (±7.0%) i/s - 12098394 in 5.013982s
#
# Comparison:
# String (Amortized): 2856024.9 i/s
# StringIO (Amortized): 2424863.3 i/s - 1.18x slower
# Adjust N to consider the cost of allocation vs the cost of appending.
N = 5
Benchmark.ips do |x|
# These next two tests consider that multiple writes may be done per buffer allocation.
x.report("String") do |i|
i.times do
buffer = String.new
N.times do
buffer << "String #{i}"
end
end
end
x.report("StringIO") do |i|
i.times do
buffer = StringIO.new
N.times do
buffer << "String #{i}"
end
end
end
x.compare!
end
# Calculating -------------------------------------
# String 36822 i/100ms
# StringIO 32471 i/100ms
# -------------------------------------------------
# String 445143.2 (±5.0%) i/s - 2246142 in 5.059017s
# StringIO 328469.2 (±4.1%) i/s - 1656021 in 5.049919s
#
# Comparison:
# String: 445143.2 i/s
# StringIO: 328469.2 i/s - 1.36x slower
Version data entries
18 entries across 18 versions & 1 rubygems