an efficient wait free resizable hash table
play

An Efficient Wait-free Resizable Hash Table Panagiota Fatourou 1,2 , - PowerPoint PPT Presentation

Sep 10, 2023 •219 likes •741 views

An Efficient Wait-free Resizable Hash Table Panagiota Fatourou 1,2 , Nikolaos Kallimanis 1 , Thomas Ropars 3 1 FORTH ICS 2 University of Crete 3 Univ. Grenoble Alpes 1 2018 A new combination of properties for a hash table Context Dictionary of

  1. An Efficient Wait-free Resizable Hash Table Panagiota Fatourou 1,2 , Nikolaos Kallimanis 1 , Thomas Ropars 3 1 FORTH ICS 2 University of Crete 3 Univ. Grenoble Alpes 1 2018

  2. A new combination of properties for a hash table Context Dictionary of Key-Value pairs Important data structure in several domains (OS, etc.) A resizable hash table Provides the strongest progress guarantee (wait-freedom) Targets the most common load for a hash table ◮ Large majority of Lookup operations Outperforms existing non-blocking algorithms for such workloads ◮ By enforcing 2 important design rules 2 2018

  3. Hash tables 00 000010 A hash function associates items to buckets 01 ◮ Fixed-size buckets 010000 00 011110 01 3 operations: 10 10 ◮ Insert(K, V) (If K 11 already exists, V is updated) ◮ Delete(K) ◮ Lookup(K) 11 3 2018

  4. Dynamic hashing 00 000010 000 010 001 010000 01 01 010011 010 010000 011 011 011110 Adapts the number of buckets 011110 100 to the number of items 101 1 Ensures constant average time 110 for operations 111 Insert(010011) 4 2018

  5. Dynamic hashing 00 000010 000 010 001 010000 01 01 010011 010 010000 011 011 011110 Adapts the number of buckets 011110 100 to the number of items 101 1 Ensures constant average time 110 for operations 111 Insert(010011) 4 2018

  6. Dynamic hashing 00 000010 000 010 001 010000 01 01 010011 010 010000 011 011 011110 Adapts the number of buckets 011110 100 to the number of items 101 1 Ensures constant average time 110 for operations 111 Insert(010011) 4 2018

  7. Extendible hashing 00 00 00 0010 000010 00 0 000 010 00 1 001 010000 Hash keys manipulated as bit 010011 010 010 strings 011 011 011 011110 ◮ A prefix of the key is used to 100 100 101 101 find the appropriate bucket 1 110 110 111 111 Resizing actions are local ◮ Splitting and merging buckets Insert(010011) 5 2018

  8. Extendible hashing 00 00 00 0010 000010 00 0 000 010 00 1 001 010000 Hash keys manipulated as bit 010011 010 010 strings 011 011 011 011110 ◮ A prefix of the key is used to 100 100 101 101 find the appropriate bucket 1 110 110 111 111 Resizing actions are local ◮ Splitting and merging buckets Insert(010011) 5 2018

  9. A wait-free concurrent hash table 00 Natural parallelism 000010 Operations applying to different 000 010 parts of the hash table can run 001 010000 in parallel 010011 010 More complex with dynamic 011 011 011110 hashing 100 101 1 110 111 T A : Insert(100000) T B : Insert(010011) 6 2018

  10. A wait-free concurrent hash table 00 Natural parallelism 000010 Operations applying to different 000 010 parts of the hash table can run 001 010000 in parallel 010011 010 More complex with dynamic 011 011 011110 hashing 100 101 1 110 Non-blocking algorithm 111 Lock freedom: At least one thread makes progress T A : Insert(100000) Wait freedom : Every T B : Insert(010011) operation completes in a finite number of steps 6 2018

  11. Towards an efficient resizable hash table: Insights Most common load for a hash table Large majority of Lookup() operations Resizing actions are rare Design rules to achieve best performance Lookup() operations should always be allowed to proceed without any synchronization When no resizing actions are executed, update operations applying to different buckets should be allowed to progress fully in parallel 7 2018

  12. Related work The split-ordered list (LF-Split) Shalev and Shavit [PODC’03] LF-Split does not comply with our design rules ◮ During Lookup() operations, threads have to help removing items marked for deletion. ◮ A global counter is modified after every insertion/deletion. LF/WF-Freeze Liu, Zhang, and Spear [PODC’14] WF-Freeze does not comply with our design rules ◮ A global sequence number is required to tag update operations 8 2018

  13. Contributions The design of a wait-free extendible hash table Follows our two design rules First algorithm to use several instances of the PSim universal construction [SPAA’11]. ◮ Appropriatly synchronized to ensure wait-freedom Experiments demonstrate the new performance trade-off Outperforms all existing non-blocking resizable hash tables when resizing actions are rare Slower resizing 9 2018

  14. Our Wait-Free Algorithm

  15. The PSim algorithm Fatourou and Kallimanis [SPAA’11] T 2 : Insert(001110) BState BState 00 00 00 b 000010 001110 000010 help: op 0 op 0 op 2 a: 0 0 0 0 a: 0 1 0 0 0 1 0 0 t: 1 1 0 0 1 0 0 0 res: res: 11 2018

  16. The PSim algorithm Fatourou and Kallimanis [SPAA’11] Announce the operation to be executed T 2 : Insert(001110) BState BState 00 00 00 b 000010 001110 000010 help: op 0 op 0 op 2 a: 0 0 0 0 a: 0 1 0 0 0 1 0 0 t: 1 1 0 0 1 0 0 0 res: res: 11 2018

  17. The PSim algorithm Fatourou and Kallimanis [SPAA’11] Announce the operation to be executed Make a local copy of the object to update T 2 : Insert(001110) BState BState 00 00 00 b 000010 001110 000010 help: op 0 op 0 op 2 a: 0 0 0 0 a: 0 1 0 0 0 1 0 0 t: 1 1 0 0 1 0 0 0 res: res: 11 2018

  18. The PSim algorithm Fatourou and Kallimanis [SPAA’11] Announce the operation to be executed Make a local copy of the object to update Apply all pending operations on the local object T 2 : Insert(001110) BState BState 00 00 00 b 000010 001110 000010 help: op 0 op 0 op 2 a: 0 0 0 0 a: 0 1 0 0 0 1 0 0 t: 1 1 0 0 1 0 0 0 res: res: 11 2018

  19. The PSim algorithm Fatourou and Kallimanis [SPAA’11] Announce the operation to be executed Make a local copy of the object to update Apply all pending operations on the local object T 2 : Insert(001110) BState BState 00 00 00 b 000010 001110 000010 help: op 0 op 0 op 2 a: 0 0 0 0 a: 0 1 0 0 0 1 0 0 t: 1 1 0 0 1 0 0 0 res: res: 11 2018

  20. The PSim algorithm Fatourou and Kallimanis [SPAA’11] Announce the operation to be executed Make a local copy of the object to update Apply all pending operations on the local object Try making the object globally visible using CAS T 2 : Insert(001110) BState BState 00 00 00 b 000010 000010 001110 help: op 0 op 0 op 2 a: 0 0 0 0 a: 1 0 0 0 1 0 0 0 t: 1 1 0 0 1 0 0 0 res: res: 11 2018

  21. The PSim algorithm Fatourou and Kallimanis [SPAA’11] Announce the operation to be executed for k in 1..2: Make a local copy of the object to update Apply all pending operations on the local object Try making the object globally visible using CAS T 2 : Insert(001110) BState BState 00 00 00 b 000010 000010 001110 help: op 0 op 0 op 2 a: 0 0 0 0 a: 1 0 0 0 1 0 0 0 t: 1 1 0 0 1 0 0 0 res: res: 11 2018

  22. The hash table structure help: ht prefix =0 00 000010 01 prefix =1 10 100000 11 DState Bucket BState Two levels of indirection One instance of PSim for the DState and for each BState 12 2018

  23. Insert (no resizing) and Lookup operations ht prefix =0 00 000010 01 prefix =1 10 100000 11 T a : Insert(111100) 100000 111100 T b : Lookup(100010) Lookup operations are executed without any synchronization ( BState objects are immutable) Insert operations on different buckets do not synchronize 13 2018

  24. Insert (no resizing) and Lookup operations ht prefix =0 00 000010 01 prefix =1 10 100000 11 T a : Insert(111100) 100000 111100 T b : Lookup(100010) Lookup operations are executed without any synchronization ( BState objects are immutable) Insert operations on different buckets do not synchronize 13 2018

  25. Insert (no resizing) and Lookup operations ht prefix =0 00 000010 01 prefix =1 10 100000 11 T a : Insert(111100) 100000 111100 T b : Lookup(100010) Lookup operations are executed without any synchronization ( BState objects are immutable) Insert operations on different buckets do not synchronize 13 2018

  26. Insert (no resizing) and Lookup operations ht prefix =0 00 000010 01 prefix =1 10 100000 11 T a : Insert(111100) 100000 111100 T b : Lookup(100010) Lookup operations are executed without any synchronization ( BState objects are immutable) Insert operations on different buckets do not synchronize 13 2018

  27. Splitting a bucket ht prefix =0 00 00 000010 01 01 prefix =1 10 10 100000 11 11 111100 00 01 prefix =10 T a : Insert(110011) 100000 10 11 prefix =11 111100 110011 14 2018

  28. Splitting a bucket ht prefix =0 00 00 000010 01 01 prefix =1 10 10 100000 11 11 111100 00 01 prefix =10 T a : Insert(110011) 100000 10 11 prefix =11 111100 110011 14 2018

  29. Splitting a bucket ht prefix =0 00 00 000010 01 01 prefix =1 10 10 100000 11 11 111100 00 01 prefix =10 T a : Insert(110011) 100000 10 11 prefix =11 111100 110011 14 2018

  30. Splitting a bucket ht prefix =0 00 00 000010 01 01 prefix =1 10 10 100000 11 11 111100 00 01 prefix =10 T a : Insert(110011) 100000 10 11 prefix =11 111100 110011 14 2018

  31. Splitting a bucket ht prefix =0 00 00 000010 01 01 prefix =1 10 10 100000 11 11 111100 00 01 prefix =10 T a : Insert(110011) 100000 10 11 prefix =11 111100 110011 To avoid losing updates: Only full buckets can be replaced during resizing No update operation can be run on a full bucket 14 2018

Recommend

Lock-Free, Wait-Free and Multi-core Programming Roger Deran boilerbay.com Fast, Efficient

Lock-Free, Wait-Free and Multi-core Programming Roger Deran boilerbay.com Fast, Efficient

Lock-Free, Wait-Free and Multi-core Programming Roger Deran boilerbay.com Fast, Efficient Concurrent Maps AirMap Lock-Free and Wait-Free Data Structures Overview The Java Maps that use Lock-Free techniques Graphical performance

538 views • 31 slides
Hash Tables Outline Overview Implementation style for the Table ADT that is Definition

Hash Tables Outline Overview Implementation style for the Table ADT that is Definition

Hash Tables Outline Overview Implementation style for the Table ADT that is Definition good in a wide range of situations is the hash Hash functions table Open hashing efficient Insert, Delete, and Search operations

321 views • 5 slides
Hash Table In a hash table, we allocate an array of size m, which is much smaller than |U|

Hash Table In a hash table, we allocate an array of size m, which is much smaller than |U|

Hash Table In a hash table, we allocate an array of size m, which is much smaller than |U| (the set of keys). Tirgul 9 We use a hash function h() to determine the entry of each key. Hash Tables (continued) Reminder The crucial

151 views • 4 slides
Resizable, Scalable, Concurrent Hash Tables via Relativistic Programming Josh Triplett 1 Paul E.

Resizable, Scalable, Concurrent Hash Tables via Relativistic Programming Josh Triplett 1 Paul E.

Resizable, Scalable, Concurrent Hash Tables via Relativistic Programming Josh Triplett 1 Paul E. McKenney 2 Jonathan Walpole 1 1 Portland State University 2 IBM Linux Technology Center June 16, 2011 Synchronization = Waiting Concurrent

989 views • 76 slides
Competitive Freshness Algorithms for Wait free Objects Wait-free Objects Peter Damaschke, Phuong

Competitive Freshness Algorithms for Wait free Objects Wait-free Objects Peter Damaschke, Phuong

Competitive Freshness Algorithms for Wait free Objects Wait-free Objects Peter Damaschke, Phuong Ha & Philippas Tsigas Presentation at the Euro-Par 2006 29 th Aug. 1 st Sept. 2006, Dresden, Germany. Introduction Wait-free data objects

365 views • 17 slides
A Lock-Free Dynamically Resizable Array Damian Dechev 1 Peter Pirkelbauer 1 Bjarne Stroustrup 1 , 2

A Lock-Free Dynamically Resizable Array Damian Dechev 1 Peter Pirkelbauer 1 Bjarne Stroustrup 1 , 2

A Lock-Free Dynamically Resizable Array Damian Dechev 1 Peter Pirkelbauer 1 Bjarne Stroustrup 1 , 2 1 Department of Computer Science Texas A&M University 2 AT&T Research OPODIS, December 2006 Parasol Lab (Texas A&M) Lock-Free Dynamic

554 views • 18 slides
Distributed Hash Tables What is a DHT? Hash Table data structure that maps keys to

Distributed Hash Tables What is a DHT? Hash Table data structure that maps keys to

Distributed Hash Tables What is a DHT? Hash Table data structure that maps keys to values essen=al building block in so?ware systems Distributed Hash Table (DHT) similar, but spread across many hosts Interface

455 views • 32 slides
Hash Table Analysis When do hash tables degrade in performance? How should we set the maximum

Hash Table Analysis When do hash tables degrade in performance? How should we set the maximum

Hash Table Analysis When do hash tables degrade in performance? How should we set the maximum load factor? It is especially important to know the average behavior of a hashing method, because we are committed to trusting in the laws of

579 views • 22 slides
From Lock-Free to Wait-Free: Linked List Edward Duong Outline 1) Outline operations of the

From Lock-Free to Wait-Free: Linked List Edward Duong Outline 1) Outline operations of the

From Lock-Free to Wait-Free: Linked List Edward Duong Outline 1) Outline operations of the locality conscious linked list [Braginsky 2012] 2) Transformation concept from lock-free -> wait-free [Timnat 2014] Why Build Concurrent ADTs

513 views • 37 slides
Hash Tables 1 Hash Table in Primary Storage Main parameter B = number of buckets Hash

Hash Tables 1 Hash Table in Primary Storage Main parameter B = number of buckets Hash

Hash Tables 1 Hash Table in Primary Storage Main parameter B = number of buckets Hash function h maps key to numbers from 0 to B-1 Bucket array indexed from 0 to B-1 Each bucket contains exactly one value Strategy

460 views • 33 slides
Chapter 6 Hash-Based Indexing Efficient Support for Equality Search Hash-Based Indexing Static

Chapter 6 Hash-Based Indexing Efficient Support for Equality Search Hash-Based Indexing Static

Hash-Based Indexing Torsten Grust Chapter 6 Hash-Based Indexing Efficient Support for Equality Search Hash-Based Indexing Static Hashing Hash Functions Architecture and Implementation of Database Systems Extendible Hashing Summer 2016

930 views • 39 slides
Cache misses for lookup, existing of random ints Cache misses for lookup, non-existing of random

Cache misses for lookup, existing of random ints Cache misses for lookup, non-existing of random

The hash tables Googles dense and sparse hash tables Use open addressing Quadratic probing SGI Is a chained hash table Referred to as gnu in graphs One-table and Two-table Doubly linked: Rather large compartments

645 views • 29 slides
Complex Libraries Using Hash Dictionaries 1 Playing Hash Table You are the new produce manager

Complex Libraries Using Hash Dictionaries 1 Playing Hash Table You are the new produce manager

Complex Libraries Using Hash Dictionaries 1 Playing Hash Table You are the new produce manager of the local grocery store. You want to use a dictionary to track your fruit inventory. Entries have the form (banana, 20) where o

595 views • 55 slides
14. Hashing Gloal: Efficient management of a table of all n ETH-students of Possible Requirement:

14. Hashing Gloal: Efficient management of a table of all n ETH-students of Possible Requirement:

Motivating Example 14. Hashing Gloal: Efficient management of a table of all n ETH-students of Possible Requirement: fast access (insertion, removal, find) of a Hash Tables, Pre-Hashing, Hashing, Resolving Collisions using dataset by name

158 views • 15 slides
CS 225 Data Structures Oc October 24 Ha Hashing G G Carl Evans A A Hash Table based

CS 225 Data Structures Oc October 24 Ha Hashing G G Carl Evans A A Hash Table based

CS 225 Data Structures Oc October 24 Ha Hashing G G Carl Evans A A Hash Table based Dictionary Client Code: 1 Dictionary<KeyType, ValueType> d; 2 d[k] = v; A Hash Table consists of three things: 1. A hash function, f(k) 2.

343 views • 20 slides
A Parallel Compact Hash Table Alfons Laarman & Steven van der Vegt Overview Research

A Parallel Compact Hash Table Alfons Laarman & Steven van der Vegt Overview Research

A Parallel Compact Hash Table Alfons Laarman & Steven van der Vegt Overview Research Motivation Background Contribution A Parallel Compact Hash Table October 3, 2011 2 / 19 Introduction Hash tables are fundamental data structures A

636 views • 29 slides
Verifying a hash table and its iterators in higher-order separation logic Franois Pottier

Verifying a hash table and its iterators in higher-order separation logic Franois Pottier

Verifying a hash table and its iterators in higher-order separation logic Franois Pottier Vocal meeting Paris, November 10, 2016 Motivation Why verify a hash table implementation ? a simple and useful data structure implements an

499 views • 19 slides
Abelian Square-Free Dithering and Recoding for Iterated Hash Functions Ronald L. Rivest MIT

Abelian Square-Free Dithering and Recoding for Iterated Hash Functions Ronald L. Rivest MIT

Abelian Square-Free Dithering and Recoding for Iterated Hash Functions Ronald L. Rivest MIT CSAIL ECRYPT Hash Function Conference June 23, 2005 Outline Dean/Kelsey/Schneier Attacks Square-Free Sequences Prouhet-Thue-Morse

402 views • 22 slides
CompSci 514: Computer Networks Lecture 13: Distributed Hash Table Xiaowei Yang Overview

CompSci 514: Computer Networks Lecture 13: Distributed Hash Table Xiaowei Yang Overview

CompSci 514: Computer Networks Lecture 13: Distributed Hash Table Xiaowei Yang Overview What problems do DHTs solve? How are DHTs implemented? Background A hash table is a data structure that stores (key, object) pairs.

580 views • 39 slides
A Sybil-Proof Distributed Hash Table Chris Lesniewski-Laas M. Frans Kaashoek MIT 28 April 2010

A Sybil-Proof Distributed Hash Table Chris Lesniewski-Laas M. Frans Kaashoek MIT 28 April 2010

A Sybil-Proof Distributed Hash Table Chris Lesniewski-Laas M. Frans Kaashoek MIT 28 April 2010 NSDI http://pdos.csail.mit.edu/whanau/slides.pptx Distributed Hash Table Interface: PUT( key , value ), GET( key ) value Route to peer

424 views • 30 slides
Ken Birman i Cornell University. CS5410 Fall 2008. What is a Distributed Hash Table (DHT)?

Ken Birman i Cornell University. CS5410 Fall 2008. What is a Distributed Hash Table (DHT)?

Ken Birman i Cornell University. CS5410 Fall 2008. What is a Distributed Hash Table (DHT)? Exactly that A service, distributed over multiple machines, with h hash table semantics h bl i Put (key, value), Value = Get (key)

849 views • 65 slides
CS200: Hash Tables Prichard Ch. 13.2 CS200 - Hash Tables 1 Table Implementations: average

CS200: Hash Tables Prichard Ch. 13.2 CS200 - Hash Tables 1 Table Implementations: average

CS200: Hash Tables Prichard Ch. 13.2 CS200 - Hash Tables 1 Table Implementations: average cases Search Add Remove Sorted O(log n) O(n) O(n) array-based Unsorted O(n) O(1) O(n) array-based Balanced O(log n) O(log n) O(log n)

540 views • 39 slides
How large should a hash CS 5633 -- Spring 2005 table be? Goal: Make the table as small as

How large should a hash CS 5633 -- Spring 2005 table be? Goal: Make the table as small as

How large should a hash CS 5633 -- Spring 2005 table be? Goal: Make the table as small as possible, but large enough so that it wont overflow (or otherwise become inefficient). Problem: What if we dont know the proper size in advance?

341 views • 7 slides
and Hash Coding with Deep Neural Networks Presenter: MinKu Kang 1 Why did I choose this paper?

and Hash Coding with Deep Neural Networks Presenter: MinKu Kang 1 Why did I choose this paper?

Simultaneous Feature Learning and Hash Coding with Deep Neural Networks Presenter: MinKu Kang 1 Why did I choose this paper? Efficient image retrieval via binary encoding of images - efficient bitwise operations. - space-efficient

606 views • 36 slides

More recommend