Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming.threads > #2226

About the bakery algorithm and concurrent FIFO queue...

Cross-posted to 2 groups.

Show all headers | View raw

Hello,


I have studied deeply the concurrent FIFO queue using the bakery 
algorithm of Chriss Thomasson, and i have wrote it in Object Pascal
and i have done all the alignment and the cache padding etc.
and i have noticed that it scored 4.8 millions of transactions
on the pop() side on my 2.4 GHz Quadcore , so even if it scored more on 
the push() side, in a scenario of high contention the throughput will be 
limited by the throughput of pop() side, so the Chriss Thomasson
conurrent FIFO queue is limited to 4.8 millions of transactions in high 
contention on my 2.4 GHz Quadcore... but here is the great news: i have
added to this concurrent FIFO queue my SemaMonitor using a simple 
spinlock with a backoff an it has scored this time 6.4 millions per 
second on the pop() side on my 2.4 GHz Quadcore, and that's more than 
the Chriss Thomasson concurrent FIFO queue that uses the bakery 
algorithm...

So i think that this fast concurrent FIFO queue that uses my SemaMonitor 
is still great and useful...

You can download all those concurrent FIFO queues from my website:

http://pages.videotron.com/aminer/




Thank you,
Amine Moulay Ramdane.

Back to comp.programming.threads | Previous | Next — Next in thread | Find similar | Unroll thread

Thread

About the bakery algorithm and concurrent FIFO queue... aminer <aminer@toto.net> - 2014-04-23 15:17 -0700
  Re: About the bakery algorithm and concurrent FIFO queue... aminer <aminer@toto.net> - 2014-04-23 15:44 -0700

csiph-web