[<prev] [next>] [day] [month] [year] [list]
Message-ID: <1696204792.18122.1472071185916.JavaMail.zimbra@efficios.com>
Date: Wed, 24 Aug 2016 20:39:45 +0000 (UTC)
From: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To: Matthew Wilcox <mawilcox@...rosoft.com>
Cc: "Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>,
linux-kernel <linux-kernel@...r.kernel.org>, rp@...s.cs.pdx.edu
Subject: RCU Judy Array (optimisation over Radix Tree)
Hi Matthew,
Following our discussion at LinuxCon about Radix Trees and Judy
Array, here is a pointer to my user-space implementation
prototype of RCU Judy Array. I've simplified it a bit (removing
features you don't need, just keeping fixed-sized integer keys),
and it can still be simplified quite a bit by removing partitioned
locking support, but it may contain nice ideas for radix tree
improvements regarding the number of cache-lines touched by radix
tree lookups.
It's available in this branch:
https://github.com/urcu/userspace-rcu/tree/urcu/rcuja-simple-int
Here are Linux Plumbers 2012 slides and video from a few years ago:
http://www.efficios.com/lpc2012-scaling-rcu-judy-arrays-cache-efficient-compact-fast-and-scalable-trie
Feedback is welcome!
Thanks,
Mathieu
--
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com
Powered by blists - more mailing lists