Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.programming.threads > #2142 > unrolled thread
| Started by | aminer <aminer@toto.net> |
|---|---|
| First post | 2014-04-11 18:53 -0700 |
| Last post | 2014-04-12 12:31 -0700 |
| Articles | 2 — 2 participants |
Back to article view | Back to comp.programming.threads
About my new scalable FIFO fair lock aminer <aminer@toto.net> - 2014-04-11 18:53 -0700
Re: About my new scalable FIFO fair lock "Chris M. Thomasson" <no@spam.invalid> - 2014-04-12 12:31 -0700
| From | aminer <aminer@toto.net> |
|---|---|
| Date | 2014-04-11 18:53 -0700 |
| Subject | About my new scalable FIFO fair lock |
| Message-ID | <li9rmu$lgk$1@news.albasani.net> |
Hello, I have just corrected a bug in my new scalable FIFO fair lock, and i have tested it thoroughly and it is stable now, other than that as you have noticed i am using a FIFO queue that uses a Ticket spinlock with a proportional, but to optimize it more for speed you can use a waitfree queue. Why have i invented my new scalable FIFO fair lock and have i not used just a TicketSpinlock with a proportional backcoff ? cause wioth the TicketSpinlock if the time in the critical section becomes bigger then the proportional backoff will not be able to do its job correctly and the TicketSpinlock will not scale to more and more cores... You can download my new scalable FIFO fair lock version 1.01 from: http://pages.videotron.com/aminer/ Thank you, Amine Moulay Ramdane.
[toc] | [next] | [standalone]
| From | "Chris M. Thomasson" <no@spam.invalid> |
|---|---|
| Date | 2014-04-12 12:31 -0700 |
| Message-ID | <lic4ba$qfu$1@speranza.aioe.org> |
| In reply to | #2142 |
> "aminer" wrote in message news:li9rmu$lgk$1@news.albasani.net... > Hello, > I have just corrected a bug in my new scalable FIFO fair lock, > and i have tested it thoroughly and it is stable now, [...] How does it compare to this: http://webee.technion.ac.il/~idish/ftp/spaa049-gidron.pdf
[toc] | [prev] | [standalone]
Back to top | Article view | comp.programming.threads
csiph-web