rabbit-slide-kou-rubykaigi-takeout-2021 2021.9.11.0 → 2021.9.11.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4e7819063a4ebcbedeb7a6ea8ffad85ca4e15a57ce9aa6dbcecd02f338610d7c
-  data.tar.gz: 6a9734a5a02321f2ce4c1e20676556c0c6efa9aa18649cfa263d1dd3832ce702
+  metadata.gz: 5d4e14fc04effc4f025a441319f3e66fe5abd2debdadf30c60018a42747a7f52
+  data.tar.gz: 60c90b5cca89e7c4a4927b5d3266b3a1cce3e4d9ab05a0ad4759b0105117c35f
 SHA512:
-  metadata.gz: a89c84882adff989423129df414b42cdc4e3f49bc61fb2651bb722105d1a4d06aa5271892cfcb790cc25a54cb6de2d11e1f335977fcf7ba59c436c97d2002eb8
-  data.tar.gz: 8a2d3ea6e38fda7c3972a8f31f5d5dd75c38ac94d852968133f27f1c99e24e0a7aa6a26155f2be330660a1012433310ca0d52bf3fe76395414cfa18b571f269c
+  metadata.gz: 9ef2cdeea3f7b003dbbbac777c1011ff4ff85bd4547790140e68e7e89728df51ac8dbe835f1d41a5589ff1868783b76327696f09af0123b56d5150ecc4abb0c6
+  data.tar.gz: be5b7964361a8d3bf1324ebcfc4cedf7a7723830faee85043a55ecbbca37eca7091159cad9ea5703dbe1d9395e67132c1afb45743ad4d575629c2a6332247b10

data/Rakefile

@@ -5,8 +5,8 @@ require "rabbit/task/slide"
 spec = nil
 Rabbit::Task::Slide.new do |task|
   spec = task.spec
-  # spec.files += Dir.glob("doc/**/*.*")
   spec.files += Dir.glob("images/**/*.*")
+  spec.files += Dir.glob("benchmarks/*.rb")
   # spec.files -= Dir.glob("private/**/*.*")
   spec.add_runtime_dependency("rabbit-theme-clear-code")
 end

data/benchmarks/add.rb

@@ -0,0 +1,37 @@
+#!/usr/bin/env ruby
+
+# Rehearsal ------------------------------------------------
+# Array          0.040361   0.004003   0.044364 (  0.044364)
+# Arrow::Array   0.004211   0.000001   0.004212 (  0.023220)
+# Numo::UInt8    0.000217   0.000001   0.000218 (  0.000217)
+# --------------------------------------- total: 0.048794sec
+#
+#                    user     system      total        real
+# Array          0.042996   0.000000   0.042996 (  0.042996)
+# Arrow::Array   0.001329   0.000000   0.001329 (  0.001329)
+# Numo::UInt8    0.000108   0.000000   0.000108 (  0.000106)
+
+require "benchmark"
+
+require "arrow"
+require "numo/narray"
+
+array = 1000000.times.to_a
+arrow_array = Arrow::Array.new(array)
+numo_array = Numo::UInt8[array]
+
+Benchmark.bmbm do |job|
+  job.report("Array") do
+    array.collect do |value|
+      value + 1
+    end
+  end
+
+  job.report("Arrow::Array") do
+    Arrow::Function.find("add").execute([arrow_array, 1]).value
+  end
+
+  job.report("Numo::UInt8") do
+    numo_array + 1
+  end
+end

data/benchmarks/filter.rb

@@ -0,0 +1,44 @@
+#!/usr/bin/env ruby
+
+# Rehearsal ------------------------------------------
+# Ruby     1.254892   0.031662   1.286554 (  1.286584)
+# Arrow    0.002342   0.000113   0.002455 (  0.002430)
+# DuckDB   0.459440   0.000000   0.459440 (  0.459448)
+# --------------------------------- total: 1.748449sec
+#
+#              user     system      total        real
+# Ruby     1.252897   0.003972   1.256869 (  1.256864)
+# Arrow    0.001361   0.000037   0.001398 (  0.001395)
+# DuckDB   0.459227   0.000000   0.459227 (  0.459101)
+
+require "benchmark"
+
+require "datasets-arrow"
+require "arrow-duckdb"
+
+dataset = Datasets::PostalCodeJapan.new
+arrow_dataset = dataset.to_arrow
+db = DuckDB::Database.open
+connection = db.connect
+connection.register("codes", arrow_dataset)
+
+Benchmark.bmbm do |job|
+  job.report("Ruby") do
+    dataset.find_all do |row|
+      row.prefecture == "東京都"
+    end
+  end
+
+  job.report("Arrow") do
+    arrow_dataset.slice do |slicer|
+      slicer.prefecture == "東京都"
+    end
+  end
+
+  job.report("DuckDB") do
+    result = connection.query("SELECT * FROM codes WHERE prefecture = ?",
+                              "東京都",
+                              output: :arrow)
+    result.to_table
+  end
+end

data/benchmarks/load.rb

@@ -0,0 +1,65 @@
+#!/usr/bin/env ruby
+
+# Rehearsal -----------------------------------------------------
+# CSV: Ruby: 0        0.867305   0.043896   0.911201 (  0.911251)
+# CSV: Red Arrow: 0   0.168953   0.044898   0.213851 (  0.016073)
+# Apache Arrow: 0     0.000396   0.000040   0.000436 (  0.000437)
+# CSV: Ruby: 1        1.820866   0.057001   1.877867 (  1.874802)
+# CSV: Red Arrow: 1   0.358689   0.055034   0.413723 (  0.021555)
+# Apache Arrow: 1     0.000393   0.000028   0.000421 (  0.000421)
+# CSV: Ruby: 2        4.141544   0.089525   4.231069 (  4.215467)
+# CSV: Red Arrow: 2   0.704372   0.133196   0.837568 (  0.040835)
+# Apache Arrow: 2     0.000447   0.000025   0.000472 (  0.000472)
+# CSV: Ruby: 3        7.919982   0.188867   8.108849 (  8.082351)
+# CSV: Red Arrow: 3   1.352553   0.235089   1.587642 (  0.080943)
+# Apache Arrow: 3     0.000544   0.000028   0.000572 (  0.000573)
+# CSV: Ruby: 4       20.722920   0.449558  21.172478 ( 21.123315)
+# CSV: Red Arrow: 4   2.570883   0.427440   2.998323 (  0.146126)
+# Apache Arrow: 4     0.000898   0.000000   0.000898 (  0.000899)
+# ------------------------------------------- total: 42.355370sec
+#
+#                         user     system      total        real
+# CSV: Ruby: 0        0.792350   0.104316   0.896666 (  0.828678)
+# CSV: Red Arrow: 0   0.173136   0.007682   0.180818 (  0.012443)
+# Apache Arrow: 0     0.000574   0.000026   0.000600 (  0.000592)
+# CSV: Ruby: 1        1.816272   0.043755   1.860027 (  1.840314)
+# CSV: Red Arrow: 1   0.387305   0.058731   0.446036 (  0.021403)
+# Apache Arrow: 1     0.000662   0.000031   0.000693 (  0.000685)
+# CSV: Ruby: 2        3.755837   0.055756   3.811593 (  3.797536)
+# CSV: Red Arrow: 2   0.702905   0.145476   0.848381 (  0.040435)
+# Apache Arrow: 2     0.000833   0.000038   0.000871 (  0.000862)
+# CSV: Ruby: 3        8.003925   0.227114   8.231039 (  8.205680)
+# CSV: Red Arrow: 3   1.250248   0.320944   1.571192 (  0.074629)
+# Apache Arrow: 3     0.004287   0.000013   0.004300 (  0.004290)
+# CSV: Ruby: 4       19.384487   0.514361  19.898848 ( 19.850910)
+# CSV: Red Arrow: 4   2.586117   0.454729   3.040846 (  0.138448)
+# Apache Arrow: 4     0.008140   0.000005   0.008145 (  0.008139)
+
+require "benchmark"
+
+require "datasets-arrow"
+
+dataset = Datasets::PostalCodeJapan.new
+table = dataset.to_arrow
+n = 5
+# n.times do |i|
+#   table.save("/tmp/codes.#{i}.csv")
+#   table.save("/tmp/codes.#{i}.arrow")
+#   table = table.concatenate([table])
+# end
+
+Benchmark.bmbm do |job|
+  n.times do |i|
+    job.report("CSV: Ruby: #{i}") do
+      CSV.read("/tmp/codes.#{i}.csv").size
+    end
+
+    job.report("CSV: Red Arrow: #{i}") do
+      Arrow::Table.load("/tmp/codes.#{i}.csv")
+    end
+
+    job.report("Apache Arrow: #{i}") do
+      Arrow::Table.load("/tmp/codes.#{i}.arrow")
+    end
+  end
+end

data/benchmarks/ractor.rb

@@ -0,0 +1,46 @@
+#!/usr/bin/env ruby
+
+# Rehearsal ----------------------------------------------
+# Sequential   4.371320   0.279862   4.651182 (  4.560637)
+# Ractor       4.930339   0.494265   5.424604 (  1.454782)
+# ------------------------------------ total: 10.075786sec
+#
+#                  user     system      total        real
+# Sequential   4.434695   0.396617   4.831312 (  4.573742)
+# Ractor       4.871155   0.518646   5.389801 (  1.454987)
+
+require "benchmark"
+
+require "datasets-arrow"
+
+table = Datasets::PostalCodeJapan.new.to_arrow
+Ractor.make_shareable(table)
+
+# Show warning
+Ractor.new {}.take
+
+n_ractors = 4
+n_jobs_per_ractor = 1000
+
+Benchmark.bmbm do |job|
+  job.report("Sequential") do
+    (n_ractors * n_jobs_per_ractor).times do
+      table.slice do |slicer|
+        slicer.prefecture == "東京都"
+      end
+    end
+  end
+
+  job.report("Ractor") do
+    ractors = n_ractors.times.collect do
+      Ractor.new(table, n_jobs_per_ractor) do |t, n|
+        n.times do
+          t.slice do |slicer|
+            slicer.prefecture == "東京都"
+          end
+        end
+      end
+    end
+    ractors.each(&:take)
+  end
+end

data/benchmarks/sum.rb

@@ -0,0 +1,35 @@
+#!/usr/bin/env ruby
+
+# Rehearsal ------------------------------------------------
+# Array          0.000759   0.000000   0.000759 (  0.000754)
+# Arrow::Array   0.002483   0.000000   0.002483 (  0.021204)
+# Numo::Int8     0.000610   0.000000   0.000610 (  0.000610)
+# --------------------------------------- total: 0.003852sec
+#
+#                    user     system      total        real
+# Array          0.000763   0.000006   0.000769 (  0.000766)
+# Arrow::Array   0.000333   0.000002   0.000335 (  0.000333)
+# Numo::Int8     0.000634   0.000004   0.000638 (  0.000636)
+
+require "benchmark"
+
+require "arrow"
+require "numo/narray"
+
+array = 1000000.times.to_a
+arrow_array = Arrow::Array.new(array)
+numo_array = Numo::Int8[array]
+
+Benchmark.bmbm do |job|
+  job.report("Array") do
+    array.sum
+  end
+
+  job.report("Arrow::Array") do
+    arrow_array.sum
+  end
+
+  job.report("Numo::Int8") do
+    numo_array.sum
+  end
+end

data/config.yaml

@@ -7,14 +7,13 @@ tags:
 - ruby
 - apache_arrow
 presentation_date: 2021-09-11
-version: 2021.9.11.0
+version: 2021.9.11.2
 licenses:
 - CC-BY-SA-4.0
-slideshare_id: 
+slideshare_id: rubykaigi-takeout-2021
 speaker_deck_id: 
-ustream_id: 
 vimeo_id: 
-youtube_id: 
+youtube_id: okXiuYiP2C4
 author:
   markup_language: :rd
   name: Sutou Kouhei

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rabbit-slide-kou-rubykaigi-takeout-2021
 version: !ruby/object:Gem::Version
-  version: 2021.9.11.0
+  version: 2021.9.11.2
 platform: ruby
 authors:
 - Sutou Kouhei
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-08-23 00:00:00.000000000 Z
+date: 2021-09-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rabbit
@@ -59,6 +59,11 @@ files:
 - ".rabbit"
 - README.rd
 - Rakefile
+- benchmarks/add.rb
+- benchmarks/filter.rb
+- benchmarks/load.rb
+- benchmarks/ractor.rb
+- benchmarks/sum.rb
 - config.yaml
 - images/apache-arrow-commits-kou.png
 - images/clear-code-rubykaigi-takeout-2021-gold-sponsor.png