Subj : Re: STL Containers w/ Virtually Zero-Overhead Lock-Free Reads...
To   : (Usenet)
From : Chris Thomasson
Date : Wed Sep 28 2005 07:15 am

> Would you like to compare your approach with this library implementation?
> http://www.cs.chalmers.se/~noble/

Not comparable. All of Nobles stuff is statically sized. My algorithm tracks 
the memory that a lock-free algorithm uses. Its a tool that can be used to 
build scaleable and dynamically sized lock-free data structures, or allow 
for the lock-free traversal of various linked structures while there are 
concurrent modifications. Mutexs, Read-Write Locks, SMR or ROP like 
algorithms are better candidates for comparison. RCU algorithms would 
probably not be a fair comparison, because in RCU you have to drop all of 
your references before you exit a quiescent state; a quiescent state is 
basically a "safe-point" between "collected-regions"...

Anyway, I have to run some documentation through another party before I have 
permission to publicly post some details on how my stuff actually works. 
This should be cleared before 2006, and then I can post some pseudo-code and 
figures. The problem is there are a couple of embodiments that fall under 
the same claims, and its hard to choose which one to publish. I am not sure 
I should post the preferred embodiment, when there are others that still 
generate high performance numbers, but don't give away something too 
private...

;) 



      [ See http://www.gotw.ca/resources/clcm.htm for info about ]
      [ comp.lang.c++.moderated.    First time posters: Do this! ]

.