[concurrency-interest] Transactional memory on GCC 4.7.0, what about Java?

Sérgio Miguel Fernandes sergio.fernandes at ist.utl.pt
Wed Nov 23 12:10:48 EST 2011


Hello All,

On 23 Nov 2011, at 15:54, Andrew Haley wrote:

> On 11/23/2011 01:08 PM, √iktor Ҡlang wrote:
>> 2011/11/23 Andrew Haley <aph at redhat.com>

>> I'm just saying that making the transactions completely transparent
>> without managed references gives the impression that there is no
>> worries to be had, which is quite to the contrary.
> 
> Whoever said that?  For STM to work well transactions should be small,
> short-lived and not do I/O.  I think everyone here knows that, at
> least.

In the context of this discussion I just wanted to let you know of the following email (previously posted to trans-memory at cs.wisc.edu):

On 7 Oct 2011, at 12:09, Sérgio Miguel Fernandes wrote:

> Dear All,
> 
> We will be presenting in SPLASH Wavefront 2011 a paper that describes our approach to use a Software Transactional Memory (JVSTM) to manage transactions in an enterprise application.
> 
> If you happen to be participating in SPLASH, feel free to come and talk to us about this.
> 
> Paper: 'Strict Serializability is Harmless: A New Architecture for Enterprise Applications'
> Sérgio Miguel Fernandes and João Cachopo
> https://dspace.ist.utl.pt/bitstream/2295/1015911/2/wafr07-fernandesps.pdf


This STM has been operating daily since 2006 in a real-world object-oriented enterprise application managing not-so-small transactions (some of them read over 1 million memory locations) with more than acceptable performance vs. the previous more-traditional approach that used an ORM and database-managed transactions.

-- 
Sérgio Miguel Fernandes
Instituto Superior Técnico
Technical university of Lisbon







More information about the Concurrency-interest mailing list