【コラム】コンピュータアーキテクチャの話 (200) Dekker's Algorithmとメモリオーダリング | エンタープライズ | マイコミジャーナル
Dekker's Algorithm しかし、排他制御を実現するのにLoad Linked/Store ConditionalやTest and Setなどの特別な命令を使わない方法も存在する。次に述べるのはDekker's Algorithmという方法である。 最初はプロセサAとプロセサBの札メモリA、Bはゼロとしておき、 プロセサAが、 (1) St [A]←1 (2) Ld R1←[B] プロセサBは、 (3) St [B]←1 (4) Ld R1←[A] と、自分の札メモリに1を書き込み、その後、相手の札メモリを読む。このようにすると、これらの命令のメモリへのアクセス順序と、プロセサA、Bが読む値は、 (1)(2)(3)(4) プロセサA:0、プロセサB:1 (1)(3)(2)(4) プロセサA:1、プロセサB:1 (1)(3)(4)(2) プロセサA:1、プロセサB:1 (3)(
Non-blocking algorithm - Wikipedia
In computer science, an algorithm is called non-blocking if failure or suspension of any thread cannot cause failure or suspension of another thread;[1] for some operations, these algorithms provide a useful alternative to traditional blocking implementations. A non-blocking algorithm is lock-free if there is guaranteed system-wide progress, and wait-free if there is also guaranteed per-thread pro
Towards a Scalable Non-Blocking Coding Style（状態遷移機械の設計に関して）
Webinar: Managing the Security and Commercial Impact of Java Updates March 12, 2020 Live WebinarJoin Azul and SoftwareONE for this webinar about the security and commercial applications of Java...Read More dev.next March 24-27, 2020 Broomfield, CO, USAAzul is one of the sponsors at this event which brings together practitioners, experts, and...Read More
Computer Laboratory – Systems Research Group – NetOS: Practical lock-free data structures
Introduction Through careful design and implementation it's possible to build data structures that are safe for concurrent use without needing to manage locks or block threads. These non-blocking data structures can increase performance by allowing extra concurrency and can improve robustness by avoiding some of the problems caused by priority inversion in local settings, or machine and link failu
Scalable Reader-Writer Locks
Scalable Reader-Writer Synchronization for Shared-Memory Multiprocessors Pseudocode from article of the above name, PPoPP '91. John M. Mellor-Crummey and Michael L. Scott. The queue-based locks below perform well in tests on machines with scores of processors. Simple, non-scalable reader-preference lock. For maximum throughput on small machines. Starves writers under continuous reader load. Starva
Scalable Synchronization
Algorithms for Scalable Synchronization on Shared-Memory Multiprocessors Pseudocode from article of the above name, ACM TOCS, February 1991. John M. Mellor-Crummey and Michael L. Scott, with later additions due to (a) Craig, Landin, and Hagersten, and (b) Auslander, Edelsohn, Krieger, Rosenburg, and Wisniewski. All of these algorithms (except for the non-scalable centralized barrier) perform well
Scheduler-Conscious Synchronization
Preemption-Safe and Scheduler-Conscious Busy-Wait Synchronization Algorithms Pseudo-code from article of the above name, ACM TOCS, February 1997. Leonidas I. Kontothanassis, Robert W. Wisniewski, and Michael L. Scott. Preemption-Safe "Handshaking" Queue Lock (Queue-HS) Scheduler-Conscious "Smart Queue" Lock (Smart-Q) Preemption-Safe Ticket Lock (Ticket-PS) Scheduler-Conscious Fair Reader-Writer Lo
Maged Michael - Selected Publications
Maged Michael: Selected Publications Journal and Conference Papers Maged M. Michael Practical Lock-Free and Wait-Free LL/SC/VL Implementations Using 64-Bit CAS pdf DISC 2004 The 18th International Conference on Distributed Computing, October 2004. Earlier versions in IBM Research Reports RC23089 (Jan. 2004) and RC23220 (May 2004). Maged M. Michael Scalable Lock-Free Dynamic Memory Allocatio
Azul Systems - Cliff Click Jr.’s Blog: A Non-Blocking HashTable
A Non-Blocking HashTable March 27, 2007 I've been wrestling with concurrent algorithms again. This time, it's a Non-Blocking (concurrent, lock-free) HashTable. I've had it basically figured out for a few months now and I'm slowing widening the circle of people I expose it too. This HashTable seems too good to be true, which is why I'm trying hard to vet it's correctness before claiming victory
1
リリース、障害情報などのサービスのお知らせ
最新の人気エントリーの配信
j次のブックマーク
k前のブックマーク
lあとで読む
eコメント一覧を開く
oページを開く
Category:Concurrency control - Wikipedia
High Performance Synchronization