[ruby-core:124570] [Ruby Bug#21838] Rails seeing degradation (20% slowdown) related to Revision 079ef92b "Implement global allocatable slots and empty pages" (from Sep 5 2024)

16 Jan 2026

      Issue #21838 has been updated by peterzhu2118 (Peter Zhu).

Thank you for reporting this issue.

I debugged it and have a fix here: https://github.com/ruby/ruby/pull/15882

I distilled the reproduction down to this script:

<details>

<summary>Script</summary>

```ruby
require "bundler/inline"

gemfile do
  source "https://rubygems.org"
  gem "ruby-vips"
  gem "benchmark"
  gem "image_processing"
  gem "shrine"
  gem "sqlite3"
  gem "rails"
end

require "rails"
require "active_record"
require "benchmark"
require "sqlite3"
require "shrine/storage/file_system"

ActiveRecord::Base.establish_connection(
  adapter: "sqlite3",
  database: ":memory:"
)

ActiveRecord::Schema.define do
  create_table "albums", force: :cascade do |t|
    t.integer "artist_id", null: false
    t.text "cover_data"
    t.datetime "created_at", null: false
    t.integer "duration", default: 0, null: false
    t.string "title"
    t.integer "tracks_count", default: 0, null: false
    t.datetime "updated_at", null: false
    t.index ["artist_id"], name: "index_albums_on_artist_id"
    t.index ["duration"], name: "index_albums_on_duration"
    t.index ["title", "artist_id"], name: "index_albums_on_title_and_artist_id", unique: true
  end

  create_table "artists", force: :cascade do |t|
    t.integer "albums_count", default: 0, null: false
    t.datetime "created_at", null: false
    t.text "description"
    t.string "name"
    t.datetime "updated_at", null: false
    t.string "website"
    t.string "wikipedia"
    t.index ["description"], name: "index_artists_on_description"
    t.index ["name"], name: "index_artists_on_name", unique: true
  end

  add_foreign_key "albums", "artists"
end

Shrine.storages = {
  cache: Shrine::Storage::FileSystem.new("/tmp", prefix: "uploads/cache"),
  store: Shrine::Storage::FileSystem.new("/tmp", prefix: "uploads/store"),
}

Shrine.plugin :activerecord
Shrine.plugin :cached_attachment_data
Shrine.plugin :restore_cached_data
Shrine.plugin :derivatives
Shrine.plugin :validation
Shrine.plugin :validation_helpers
Shrine.plugin :determine_mime_type
Shrine.plugin :remote_url, max_size: 1*1024*1024

class CoverUploader < Shrine
  Attacher.derivatives do |original|
    vips = ImageProcessing::Vips.source(original)

    {
      small: vips.resize_to_limit!(300, 300),
      medium: vips.resize_to_limit!(500, 500)
    }
  end
end

class ApplicationRecord < ActiveRecord::Base
  primary_abstract_class
end

class Album < ApplicationRecord
  include CoverUploader::Attachment(:cover)

  # ASSOCIATIONS
  belongs_to :artist, counter_cache: true, touch: true

  # VALIDATIONS
  validates :title, presence: true, uniqueness: {scope: :artist_id}
end

class Artist < ApplicationRecord
  # ASSOCIATIONS
  has_many :albums, dependent: :destroy, strict_loading: true

  # VALIDATIONS
  validates :name, presence: true, uniqueness: {case_sensitive: false}

  # Only allow http:// and https:// URL schemes.
  validates :website, format: {
    with: URI::DEFAULT_PARSER.make_regexp(%w[http https]),
    allow_blank: true
  }
end

result = Benchmark.measure do
  artist_records = 500.times.map do |i|
    {
      name: "Artist #{i}",
      description: "Description for artist #{i}",
      wikipedia: "https://en.wikipedia.org/wiki/Artist_#{i}",
      website: "https://artist#{i}.com"
    }
  end
  Artist.insert_all(artist_records)

  album_records = 1000.times.map do |i|
    {
      title: "Album #{i}",
      artist_id: (i % 500) + 1,
    }
  end
  Album.insert_all(album_records)

  Shrine.storages[:cache].clear!
  Shrine.storages[:store].clear!
  image_benchmark = Benchmark.measure do
    Album.find_each do |album|
      cover_path = "/home/peter/Downloads/Cover.jpg"

      # Local cover upload.
      album.cover = File.open(cover_path)
      album.cover_derivatives! # Generate derivatives via libvips.
      album.save
    end
  end
  puts "\n=== Image Processing Only ==="
  puts "User CPU: #{image_benchmark.utime}s"
  puts "System CPU: #{image_benchmark.stime}s"
  puts "Real time: #{image_benchmark.real}s"
  puts "==============================\n"

  Shrine.storages[:cache].clear!
end

puts "\n=== Benchmark Results ==="
puts "User CPU: #{result.utime}s"
puts "System CPU: #{result.stime}s"
puts "Real time: #{result.real}s"
puts "========================\n"

puts GC.stat
pp GC.stat_heap
```

</details>

Before:

```
{count: 2017, time: 918, marking_time: 559, sweeping_time: 359, heap_allocated_pages: 333, heap_empty_pages: 0, heap_allocatable_slots: 5546, heap_available_slots: 287962, heap_live_slots: 287105, heap_free_slots: 857, heap_final_slots: 0, heap_marked_slots: 269412, heap_eden_pages: 333, total_allocated_pages: 333, total_freed_pages: 0, total_allocated_objects: 3459213, total_freed_objects: 3172108, malloc_increase_bytes: 2259851, malloc_increase_bytes_limit: 16777216, minor_gc_count: 1987, major_gc_count: 30, compact_count: 0, read_barrier_faults: 0, total_moved_objects: 0, remembered_wb_unprotected_objects: 0, remembered_wb_unprotected_objects_limit: 2684, old_objects: 269126, old_objects_limit: 536956, oldmalloc_increase_bytes: 2259851, oldmalloc_increase_bytes_limit: 16777216}
{0 => {slot_size: 40, heap_live_slots: 167011, heap_free_slots: 1639, heap_final_slots: 0, heap_eden_pages: 103, heap_eden_slots: 168650, total_allocated_pages: 119, force_major_gc_count: 3, force_incremental_marking_finish_count: 2, total_allocated_objects: 2412984, total_freed_objects: 2245973},
 1 => {slot_size: 80, heap_live_slots: 79897, heap_free_slots: 299, heap_final_slots: 0, heap_eden_pages: 98, heap_eden_slots: 80196, total_allocated_pages: 322, force_major_gc_count: 22, force_incremental_marking_finish_count: 0, total_allocated_objects: 589564, total_freed_objects: 509667},
 2 => {slot_size: 160, heap_live_slots: 33658, heap_free_slots: 273, heap_final_slots: 0, heap_eden_pages: 83, heap_eden_slots: 33931, total_allocated_pages: 189, force_major_gc_count: 3, force_incremental_marking_finish_count: 0, total_allocated_objects: 436213, total_freed_objects: 402555},
 3 => {slot_size: 320, heap_live_slots: 3306, heap_free_slots: 158, heap_final_slots: 0, heap_eden_pages: 17, heap_eden_slots: 3464, total_allocated_pages: 52, force_major_gc_count: 0, force_incremental_marking_finish_count: 0, total_allocated_objects: 15741, total_freed_objects: 12435},
 4 => {slot_size: 640, heap_live_slots: 3271, heap_free_slots: 88, heap_final_slots: 0, heap_eden_pages: 33, heap_eden_slots: 3359, total_allocated_pages: 94, force_major_gc_count: 1, force_incremental_marking_finish_count: 0, total_allocated_objects: 4749, total_freed_objects: 1478}}
```

After:

```
{count: 2017, time: 541, marking_time: 205, sweeping_time: 336, heap_allocated_pages: 336, heap_empty_pages: 0, heap_allocatable_slots: 18931, heap_available_slots: 290435, heap_live_slots: 287797, heap_free_slots: 2638, heap_final_slots: 0, heap_marked_slots: 270104, heap_eden_pages: 336, total_allocated_pages: 336, total_freed_pages: 0, total_allocated_objects: 3460006, total_freed_objects: 3172209, malloc_increase_bytes: 2259899, malloc_increase_bytes_limit: 16777216, minor_gc_count: 2009, major_gc_count: 8, compact_count: 0, read_barrier_faults: 0, total_moved_objects: 0, remembered_wb_unprotected_objects: 0, remembered_wb_unprotected_objects_limit: 2663, old_objects: 269818, old_objects_limit: 532620, oldmalloc_increase_bytes: 2259899, oldmalloc_increase_bytes_limit: 18971452}
{0 => {slot_size: 40, heap_live_slots: 167111, heap_free_slots: 1545, heap_final_slots: 0, heap_eden_pages: 103, heap_eden_slots: 168656, total_allocated_pages: 126, force_major_gc_count: 3, force_incremental_marking_finish_count: 2, total_allocated_objects: 2413543, total_freed_objects: 2246432},
 1 => {slot_size: 80, heap_live_slots: 80426, heap_free_slots: 595, heap_final_slots: 0, heap_eden_pages: 99, heap_eden_slots: 81021, total_allocated_pages: 360, force_major_gc_count: 2, force_incremental_marking_finish_count: 0, total_allocated_objects: 590103, total_freed_objects: 509677},
 2 => {slot_size: 160, heap_live_slots: 33750, heap_free_slots: 183, heap_final_slots: 0, heap_eden_pages: 83, heap_eden_slots: 33933, total_allocated_pages: 200, force_major_gc_count: 0, force_incremental_marking_finish_count: 0, total_allocated_objects: 436028, total_freed_objects: 402278},
 3 => {slot_size: 320, heap_live_slots: 3254, heap_free_slots: 5, heap_final_slots: 0, heap_eden_pages: 16, heap_eden_slots: 3259, total_allocated_pages: 33, force_major_gc_count: 0, force_incremental_marking_finish_count: 0, total_allocated_objects: 15623, total_freed_objects: 12369},
 4 => {slot_size: 640, heap_live_slots: 3294, heap_free_slots: 272, heap_final_slots: 0, heap_eden_pages: 35, heap_eden_slots: 3566, total_allocated_pages: 47, force_major_gc_count: 0, force_incremental_marking_finish_count: 0, total_allocated_objects: 4747, total_freed_objects: 1453}}
```

Note that major GC count has reduced significantly, from 30 to 8.

----------------------------------------
Bug #21838: Rails seeing degradation (20% slowdown) related to Revision 079ef92b "Implement global allocatable slots and empty pages" (from Sep 5 2024)
https://bugs.ruby-lang.org/issues/21838#change-116144

* Author: bluz71 (Dennis B)
* Status: Open
* ruby -v: ruby 3.4.0dev (2024-09-09T14:15:21Z v3_4_0_preview2~545 079ef92b5e)
* Backport: 3.2: UNKNOWN, 3.3: UNKNOWN, 3.4: UNKNOWN, 4.0: UNKNOWN
----------------------------------------
Hello,

I noticed a performance degradation with Ruby 3.4 when seeding my Rails application, which does image derivation via the Shrine Gem and `libvips` image library.

I skipped Ruby 3.4 hoping that 4.0 would magically solve my performance issue. Alas not.

My degradation when doing `rails db:reset` (aka database seeding) was about 20% as compared with Ruby 3.3.

The last couple of days I looked at this issue more closely and with the help of `git bisect` I believe this commit: https://bugs.ruby-lang.org/projects/ruby-master/repository/git/revisions/079... is the cause of my regression. That commit implements global allocatable slots and empty pages.

The revision before is fast when seeding my Rails application.
This revision however slows down my Rails seeding by about 20%, from 30 seconds to 40 seconds consistently.

I created a minimal Rails application as the reproducible test case: https://github.com/bluz71/rails_app_ruby_slowdown

Steps to reproduce:
1. `git clone https://github.com/bluz71/rails_app_ruby_slowdown`
2. `cd rails_app_ruby_slowdown`
3. `bundle config set force_ruby_platform true`
4. `bundle install`
5. `bundle exec rails db:create`
6. `RUBY_YJIT_ENABLE=0 time bundle exec rails db:reset`

Note, `libvips` needs to be installed at the system level, for example `sudo apt install libvips`, libvips is the image processing library used by Shrine which in turn is used in this Rails application when seeding.

Use Ruby 3.3 and 3.4 to compare, or use Revision de7ac11a (https://bugs.ruby-lang.org/projects/ruby-master/repository/git/revisions/de7...) and Revision 079ef92b (https://bugs.ruby-lang.org/projects/ruby-master/repository/git/revisions/079...) to compare.

I hope this is reproducible, if so, I also hope this performance loss can be minimised in future.

Best regards.

-- 
https://bugs.ruby-lang.org/