[seam-dev] Seam 3 and fork/join framework in JDK 7

[Arbi Sookazian]

I would be interested in seeing performance monitoring results for a load
test of the Seam 2.x booking app vs. Seam 3 booking app.  Using MySQL db or
similar, not HSQLDB.

Is Seam 3 in a usable state currently?  AFAIK it's pre-alpha or alpha so is
there any timeline for a M1 or beta, etc.?

On Fri, Jan 29, 2010 at 11:46 AM, Pete Muir <pmuir at redhat.com> wrote:

> Well, it largely depends on the problem set you are addressing. This kind
> of concurrency is clearly useful on multi-core systems if you are addressing
> a problem set which which needs to extensively apply functions to ordered
> lists (aka sorting).
>
> Seam 3 won't have a huge need to work with such things.
>
> Weld performs all dependency resolution and bean metadata creation at
> startup, meaning the runtime overhead of a resolution is O(1) - a hash table
> lookup. Current measurements show Weld's metadata startup time to be
> negliable compared the time taken to read class definitions - if you can
> show me otherwise, please send me instructions to reproduce in a profiler.
>
> Likewise, if you can show me hotspots at runtime, please send instructions
> to reproduce.
>
> On 29 Jan 2010, at 19:35, Arbi Sookazian wrote:
>
> > Thx for the response Pete.  I'm trying to gauge how helpful or critical
> it will be to do parallel/concurrent programming via the ParellelArray class
> (
> http://gee.cs.oswego.edu/dl/jsr166/dist/extra166ydocs/extra166y/ParallelArray.html)
> like in TNeward's example for Seam 3 apps or using a JVM language like
> Scala.  And I'm talking in a Seam 3 context only.
> >
> > In terms of identifying performance bottlenecks in Seam 2.x apps, it
> seems much of it has to do with the JSF lifecycle and Seam interceptors and
> dynamic injection (injection occurs on all @In annotated properties in a
> class whenever a public business interface method is invoked - and this will
> most likely change in Seam 3 due to the once only injection strategy in CDI
> - inject only upon instantiation of the managed bean).
> >
> > Would the fork/join frmwk or Scala help at all in terms of performance
> optimization or should we be worried more about the db tier which is
> typically the least scalable of an n-tier web system?  Or other obvious
> things like 2nd level cache implementation or avoiding/dealing with n+1
> selects problems...
> >
> > On Fri, Jan 29, 2010 at 11:20 AM, Pete Muir <pmuir at redhat.com> wrote:
> > We plan to support Java 5 and above for the foreseeable future. So yes,
> Seam 3 will be run on Java 7 (as Java SE has backwards binary compatibility)
> but won't take advantage of new features in Java 7.
> >
> > On 29 Jan 2010, at 18:50, Arbi Sookazian wrote:
> >
> > > Hi all,
> > >
> > > I just read this interesting article on fork/join frmwk by Ted Neward:
> http://www.devx.com/SpecialReports/Article/40982
> > >
> > > I'm wondering if there are any plans to use or recommend Seam 3
> developers to use the fork/join frmwk in Seam 3 if Seam 3 will run on JDK 7.
>  I'm guessing the final release dates for OpenJDK and Seam 3 may possibly be
> around the same time (some time after June this year?)
> > >
> > > Will Seam 3 be compatible with Java 7?
> > >
> > > thx.
> > > _______________________________________________
> > > seam-dev mailing list
> > > seam-dev at lists.jboss.org
> > > https://lists.jboss.org/mailman/listinfo/seam-dev
> >
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/seam-dev/attachments/20100129/997bd70e/attachment.html