Re: [abc-users] New around weaving strategy

A bit more about around inlining ...

I should have added that the classID-shadowID will still show
up in cases where inlining would be over size thresholds and
circular advice. In these cases you may find a mixture of inlining
and switches - with some cases in the switch inlined and others not.

Cheers, Laurie

+-----------------------------------------------------------------
| Laurie Hendren --- laurie.hendren@mcgill.ca
| Associate Dean (Academic), Faculty of Science,
| Dawson Hall, McGill University, 853 Sherbrooke St W,
| Montreal QC H3A 2T6 Canada, 514-398-7179, fax 514-398-1774
+----------------------------------------------------------------
| For contact and home page info as Professor, Computer Science:
| http://www.sable.mcgill.ca/~hendren --- hendren@cs.mcgill.ca
| Research: http://www.sable.mcgill.ca http://aspectbench.org
+----------------------------------------------------------------

On Mon, 17 Oct 2005, Prof. Laurie HENDREN wrote:

> Hi Eduardo,
>
> We are still working on several different around strategies. The
> inlining you see in 1.1.0 is actually an inlining of the original
> classID-shadowID weaving and you will note that the way in which it
> inlines is different than in ajc. This does sometimes give a
> performance improvement in speed, but at a penalty in space.
>
> When we have some time we will be writing a more detailed paper on
> the subject of the around weaver, and introducing yet another variation
> and studying in depth the performance of the different strategies.
>
> CHeers. Laurie
>
>
> +-----------------------------------------------------------------
> | Laurie Hendren --- laurie.hendren@mcgill.ca
> | Associate Dean (Academic), Faculty of Science,
> | Dawson Hall, McGill University, 853 Sherbrooke St W,
> | Montreal QC H3A 2T6 Canada, 514-398-7179, fax 514-398-1774
> +----------------------------------------------------------------
> | For contact and home page info as Professor, Computer Science:
> | http://www.sable.mcgill.ca/~hendren --- hendren@cs.mcgill.ca
> | Research: http://www.sable.mcgill.ca http://aspectbench.org
> +----------------------------------------------------------------
>
> On Mon, 17 Oct 2005, Eduardo S. Cordeiro wrote:
>
> > Hello abc developers,
> >
> > I've been looking at the code woven by abc version 1.1.0 and I noticed
> > that it no longer uses the class_ID - shadow_ID "routing" advice and
> > shadows described in the paper "Optimising AspectJ" and on Kuzin's
> > master thesis, but rather it inlines the advice code.
> >
> > In which cases does abc 1.1.0 prefer to inline? Is the shadow_ID
> > strategy still used in circular advice applications? Was there any
> > performance gain on your benchmarks with the inlining weaver?
> >
> > Thank you and best regards,
> >
> > Eduardo Cordeiro
> >
> >
> >
>
>
>
>
Received on Mon Oct 17 13:19:41 2005