All Projects → jtmueller → Collections.pooled

jtmueller / Collections.pooled

Licence: mit
Fast, low-allocation ports of List, Dictionary, HashSet, Stack, and Queue using ArrayPool and Span.

Projects that are alternatives of or similar to Collections.pooled

Mlib
Library of generic and type safe containers in pure C language (C99 or C11) for a wide collection of container (comparable to the C++ STL).
Stars: ✭ 321 (+179.13%)
Mutual labels:  stack, dictionary, list, queue
Learningmasteringalgorithms C
Mastering Algorithms with C 《算法精解:C语言描述》源码及Xcode工程、Linux工程
Stars: ✭ 615 (+434.78%)
Mutual labels:  stack, list, queue
Android interviews
🚀Everything you need to know to find a android job. 算法 / 面试题 / Android 知识点 🔥🔥🔥 总结不易,你的 star 是我最大的动力!
Stars: ✭ 510 (+343.48%)
Mutual labels:  stack, list, queue
Containers
This library provides various containers. Each container has utility functions to manipulate the data it holds. This is an abstraction as to not have to manually manage and reallocate memory.
Stars: ✭ 125 (+8.7%)
Mutual labels:  stack, list, queue
Redux Data Structures
Reducer factory functions for common data structures: counters, maps, lists (queues, stacks), sets, etc.
Stars: ✭ 157 (+36.52%)
Mutual labels:  stack, dictionary, queue
Buckets Js
A complete, fully tested and documented data structure library written in pure JavaScript.
Stars: ✭ 1,128 (+880.87%)
Mutual labels:  stack, dictionary, queue
Data Structures With Go
Data Structures with Go Language
Stars: ✭ 121 (+5.22%)
Mutual labels:  stack, dictionary, queue
C Macro Collections
Easy to use, header only, macro generated, generic and type-safe Data Structures in C
Stars: ✭ 192 (+66.96%)
Mutual labels:  stack, list, queue
Gostl
Data structure and algorithm library for go, designed to provide functions similar to C++ STL
Stars: ✭ 254 (+120.87%)
Mutual labels:  stack, list, queue
Javascript Datastructures Algorithms
📚 collection of JavaScript and TypeScript data structures and algorithms for education purposes. Source code bundle of JavaScript algorithms and data structures book
Stars: ✭ 3,221 (+2700.87%)
Mutual labels:  stack, dictionary, queue
Datastructure
常用数据结构及其算法的Java实现,包括但不仅限于链表、栈,队列,树,堆,图等经典数据结构及其他经典基础算法(如排序等)...
Stars: ✭ 419 (+264.35%)
Mutual labels:  stack, list, queue
Cdsa
A library of generic intrusive data structures and algorithms in ANSI C
Stars: ✭ 549 (+377.39%)
Mutual labels:  stack, queue
Qlibc
qLibc is a simple and yet powerful C library providing generic data structures and algorithms
Stars: ✭ 614 (+433.91%)
Mutual labels:  stack, queue
Coding Interview Gym
leetcode.com , algoexpert.io solutions in python and swift
Stars: ✭ 451 (+292.17%)
Mutual labels:  stack, queue
Iruka
A collection of classical data structures ⛩ and algorithms 🏃‍♂️ implemented in Typescript with video lectures 📹.
Stars: ✭ 625 (+443.48%)
Mutual labels:  stack, queue
Algorithm
Algorithm is a library of tools that is used to create intelligent applications.
Stars: ✭ 787 (+584.35%)
Mutual labels:  stack, dictionary
Cracking The Coding Interview
Solutions for Cracking the Coding Interview - 6th Edition
Stars: ✭ 35 (-69.57%)
Mutual labels:  stack, queue
Libgenerics
libgenerics is a minimalistic and generic library for C basic data structures.
Stars: ✭ 42 (-63.48%)
Mutual labels:  stack, queue
Algodeck
An Open-Source Collection of 200+ Algorithmic Flash Cards to Help you Preparing your Algorithm & Data Structure Interview 💯
Stars: ✭ 4,441 (+3761.74%)
Mutual labels:  stack, queue
Liblist
Generic Linked list Management Library in C
Stars: ✭ 22 (-80.87%)
Mutual labels:  list, queue

Collections.Pooled

Master Latest
Build status Build status

This library is based on classes from System.Collections.Generic that have been altered to take advantage of the new System.Span<T> and System.Buffers.ArrayPool<T> libraries to minimize memory allocations, improve performance, and/or allow greater interoperablity with modern API's.

Collections.Pooled supports both .NET Standard 2.0 (.NET Framework 4.6.1+) as well as an optimized build for .NET Core 2.1+. An extensive set of unit tests and benchmarks have been ported from corefx.

Total tests: 27501. Passed: 27501. Failed: 0. Skipped: 0.
Test Run Successful.
Test execution time: 9.9019 Seconds

Installation

NuGet Version

Install-Package Collections.Pooled
dotnet add package Collections.Pooled
paket add Collections.Pooled

Benchmarks

PooledList<T>

PooledList<T> is based on the corefx source code for System.Collections.Generic.List<T>, modified to use ArrayPool for internal array-storage allocation, and to support Span<T>.

There are some API changes worth noting:

  • Find and FindLast have become TryFind and TryFindLast, following the standard .NET "try" pattern.
  • The new PooledList<T>.Span property returns a Span<T> over the portion of the internal array store that is populated. This span can be further sliced, or passed into other methods that can read from or write to a span.
    • Enumerating the Span property can be more than twice as fast as enumerating the list itself, but as we are unable to do the normal "collection modified during enumeration" checks, please use with caution.
    • The list is unable to increment its version-checking flag when setting values via the Span. Therefore, "collection modified during enumeration" checks can be bypassed. Please use with caution.
  • GetRange now returns a Span<T>. For example, foo.GetRange(5, 100) is equivalent to foo.Span.Slice(5, 100).
  • CopyTo now takes a Span<T> and doesn't offer any start-index or count parameters. Use the Span property and slicing instead.
  • AddRange and InsertRange can now accept a ReadOnlySpan<T>.
  • The new AddSpan and InsertSpan methods ensure the internal storage has the capacity to add the requested number of items, and return a Span<T> that can be used to write directly to that section of the internal storage. Caveats about "collection modified during enumeration" checks apply here as well.
  • Delegate types such as Predicate<T> and Converter<T1, T2> have been replaced with standard Func<> equivalents.
  • PooledList implements IDisposable. Disposing the list returns the internal array to the ArrayPool. If you forget to dispose the list, nothing will break, but memory allocations and GC pauses will be closer to those of List<T> (you will still benefit from pooling of intermediate arrays as the PooledList is resized).
  • A selection of ToPooledList() extension methods are provided.
  • You can optionally supply a custom implementation of ArrayPool<T> to the constructor, to be used instead of the default ArrayPool<T>.Shared pool.
  • The clearMode constructor parameter gives you control over whether data is cleared before returning arrays to the ArrayPool.
  • The sizeToCapacity constructor parameter causes the list to start out with Count == Capacity. All entries in the list will have the default value for the type, or if clearMode is set to ClearMode.Never then entries in the list may have a previously-used value from the array pool. This feature is primarily useful when working with value types and avoiding unnecessary allocations.

Performance

Adding items to a list is one area where ArrayPool helps us quite a bit: List Add Benchmarks

PooledDictionary<TKey, TValue>

PooledDictionary<TKey, TValue> is based on the corefx source code for System.Collections.Generic.Dictionary<TKey, TValue>, modified to use ArrayPool for internal storage allocation, and to support Span<T>.

There are some API changes worth noting:

  • New methods include: AddRange, GetOrAdd, AddOrUpdate
  • Both constructors and AddRange can take a sequence of KeyValuePair<TKey, TValue> objects, or a sequence of ValueTuple<TKey, TValue> objects.
  • Significantly reduced memory allocations when adding many items.
  • PooledDictionary implements IDisposable. Disposing the dictionary returns the internal arrays to the ArrayPool. If you forget to dispose the dictionary, nothing will break, but memory allocations and GC pauses will be closer to those of Dictionary<TKey, TValue> (you will still benefit from pooling of intermediate arrays as the PooledDictionary is resized).
  • A selection of ToPooledDictionary() extension methods are provided.
  • The ClearMode constructor parameter gives you control over whether data is cleared before returning arrays to the ArrayPool.

Performance

Adding to dictionaries is where using ArrayPool really has an impact: Dictionary Add Benchmarks

PooledSet<T>

PooledSet<T> is based on the corefx source code for System.Generic.Collections.HashSet<T>, modified to use ArrayPool for internal storage allocation, and to support ReadOnlySpan<T> for all set functions.

  • Constructors, and all set methods have overloads that accept ReadOnlySpan<T>.
  • There are also overloads that accept arrays but forward to the span implementation. They are present to avoid ambiguous method calls.
  • PooledSet slightly outstrips HashSet in performance in almost every case, while allocating less memory. In cases where the internal arrays must be resized, PooledSet enjoys a substantial advantage.
  • PooledSet implements IDisposable. Disposing the set returns the internal arrays to the ArrayPool. If you forget to dispose the set, nothing will break, but memory allocations and GC pauses will be closer to those of HashSet<T> (you will still benefit from pooling of intermediate arrays as the PooledSet is resized).
  • A selection of ToPooledSet() extension methods are provided.
  • The ClearMode constructor parameter gives you control over whether data is cleared before returning arrays to the ArrayPool.

Performance

Here's what pooling does for us when adding to a PooledSet.

Set Add Benchmarks

PooledStack<T>

PooledStack<T> is based on the corefx source code for System.Generic.Collections.Stack<T>, modified to use ArrayPool for internal storage allocation.

  • Other than the ability to pass Spans into the constructor, the only other API change from Stack<T> is the addition of the RemoveWhere method: because sometimes you just need to remove something from a stack.
  • Significantly reduced memory allocations when pushing many items.
  • PooledStack implements IDisposable. Disposing the stack returns the internal array to the ArrayPool. If you forget to dispose the stack, nothing will break, but memory allocations and GC pauses will be closer to those of Stack<T> (you will still benefit from pooling of intermediate arrays as the PooledStack is resized).
  • A selection of ToPooledStack() extension methods are provided.
  • You can optionally supply a custom implementation of ArrayPool<T> to the constructor, to be used instead of the default ArrayPool<T>.Shared pool.
  • The ClearMode constructor parameter gives you control over whether data is cleared before returning arrays to the ArrayPool.

Performance

Once again, pushing to a stack shows off some of the advantages of using ArrayPool: Stack Push Benchmarks

PooledQueue<T>

PooledQueue<T> is based on the corefx source code for System.Generic.Collections.Queue<T>, modified to use ArrayPool for internal storage allocation.

  • Other than the ability to pass Spans into the constructor, the only other API change from Queue<T> is the addition of the RemoveWhere method: because sometimes you just need to remove something from a queue.
  • Significantly reduced memory allocations when enqueueing many items.
  • PooledQueue implements IDisposable. Disposing the queue returns the internal array to the ArrayPool. If you forget to dispose the queue, nothing will break, but memory allocations and GC pauses will be closer to those of Queue<T> (you will still benefit from pooling of intermediate arrays as the PooledQueue is resized).
  • A selection of ToPooledQueue() extension methods are provided.
  • You can optionally supply a custom implementation of ArrayPool<T> to the constructor, to be used instead of the default ArrayPool<T>.Shared pool.
  • The ClearMode constructor parameter gives you control over whether data is cleared before returning arrays to the ArrayPool.

Performance

Queue Enqueue Benchmarks

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].