harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Xiao-Feng Li" <xiaofeng...@gmail.com>
Subject Re: [drlvm][gcv5] finalizer design
Date Wed, 20 Dec 2006 01:40:13 GMT
On 12/20/06, Rana Dasgupta <rdasgupt@gmail.com> wrote:
> On 12/18/06, Xiao-Feng Li <xiaofeng.li@gmail.com> wrote:
> >
> >
> > >For "WBS", I personally think the idea of WBS is buggy. For
> > >finalization, increasing thread number to compete with Java thread for
> > >scheduling opportunity is problematic. This idea is an assumption that
> > >the underlying scheduling system in OS will work as WBS expects, but
> > >this assumption is not necessarily true.
> There is some underlying assumption of Java to OS threads being 1:1, and
> ultimately the OS scheduling everything. Some of these will need to change
> where the OS is not the hosting environment, eg., running Harmony with
> Oracle DB or Oracle App Server which will want to do their own task
> scheduling. We can expect some surprises and new data when we run under
> these environments.
> >As I said, the native finalization implementation is still evolving. I
> > >would suggest to have some benchmarks to compare the Java thread
> > >implementation and the native one, which can help us to understand the
> > >design better.
> SpecJAppServer. It's the kind of application environment we want to excel in
> anyway.

Thanks! This info is important. We will try to setup a SpecJAppServer.
We tried a couple of months ago, but then Harmony is not ready to run
it. Now it's time to have a second try.


View raw message