Subj : Re: Distributed Proxy Reference Counting: The VZOOM Algorithm...
To   : comp.programming.threads
From : Joe Seigh
Date : Mon Sep 19 2005 09:14 am

Chris Thomasson wrote:
>>I figured out that SMR and RCU are not very
>>practical for a "general solution" to the reader/writer problem. For those
>>who are interested, let me very briefly explain why:
> 
> 
> I also believe that SUN has a patent application out on something called the 
> "Repeat Offender Problem (ROP)". AFAICT it uses a DWCAS to place a "guard" 
> on a data object. It looks like SUN's algorithm may be more expensive than 
> SMR...
> 
> Any idea why SUN may be using DWCAS for this? 
> 
> 
ROP is similar to SMR in some aspects.  Sun has gone beyond DWCAS (or DCAS)
at this point.  They're into KCSS (k compare, single swap) and transactional
memory for obstruction-free (not lock-free) algorithms.

-- 
Joe Seigh

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

.