Subj : Re: Proxy collector based on RCU+SMR hazard pointers
To   : comp.programming.threads
From : Joe Seigh
Date : Sun Aug 14 2005 10:45 pm

d|dq wrote:
> Joe Seigh wrote:
> 
>> Doing a proxy collector based on RCU+SMR is easier than
>> the proxy collector based on atomic_ptr since you don't
>> need a special collector object.  Just keep a spare collection
>> element object around.  When you remove an element from the
>> collection, instead of deferring deletion of it, you swap it
>> with the proxy object and defer deletion of it instead.  The
>> built in FIFO of deletes if you're using mutexes for the writes
>> and FIFO processing of RCU+SMR take care of everything, so no
>> explicit FIFO queue like atomic_ptr proxy does.
>>
>> Performance measurements on a single cpu show RCU+SMR proxy performs
>> way better than atomic_ptr proxy which did slightly better than
>> plain RCU+SMR on linked list traversals with no preemption.
>>
> 
> What is it? :)
> 
> Please send me read me something that will tell me more about the 
> background of this information.

The RCU+SMR stuff is here
http://atomic-ptr-plus.sourceforge.net/

The atomic_ptr proxy stuff is in the atomic-ptr-plus package in
appc.h which uses atomic_ptr.  I haven't published the proxy
collector based on RCU+SMR yet.  I should probably test it a
bit more first.

-- 
Joe Seigh

When you get lemons, you make lemonade.
When you get hardware, you make software.

.