Re: [abc] pldi

I'm getting ready to re-run the experiments.

Is abc-1.0.2 the version I should be running?

Damien: Are the improvements you made to the cflow implementation in
1.0.2? Are they on by default, or do I have to turn them on with some
switch? Do we just want to run with them on, or do we also want to run
with them off for comparison?

One of the reviewers said that we were not very clear about reporting
times for no optimization, intra-only, and intra+inter. He suggested
that we also do inter-only. I spoke with Laurie, and we decided that we
should make it clearer what we did, but we shouldn't run inter-only.
On all but one benchmark, inter removes all the cflow stuff, so
inter+intra would be exactly the same code as inter-only, so there's no
point in doing inter-only.

Ondrej

On Fri, Apr 08, 2005 at 08:33:43AM -0400, Prof. Laurie HENDREN wrote:
> We need to make our final PLDI paper. I have added some files to
> the pldi05abc/ svn directory and started to update a file called
> final.tex which uses the right .cls file (they have changed the style
> file).
>
> I have also added the files PLDI05reviews.txt and
> PLDI05submissioninstructions.txt.
>
> We have two more pages (even a bit more, since the new style seems more
> squished than the old one).
>
> It is due next Friday (April 15) - even though our original notification
> said April 11, the more recent one from the publishers said April 15.
> That's good since I have been completely snowed under with the end of
> term and committees these last two weeks.
>
> Here are the things we need to do (we may have to add things to this list):
> - rerun experiments (Ondrej) ... note the reviewers wanted us to split
> up the intra and inter effects, and also look at some version
> of ajc just before and just after they added cflow counters ...
>
> - around inlining - Sascha ... did you want to play a bit with your
> more tuned inliners? Are there better experiments to do now?
> - also Chris G. noticed that after your do your inlining there are
> often dead calls to AspectOf() ... can these be removed?
>
> - cflow opts ... Damien ... at AOSD you mentionned that the interproc
> test might be stronger than necessary - I don't remember the details,
> can you comment? Should we add more to the intraproc cflow opts
> now?
>
> - other opts ... we originally had a section called "other opts" that
> we commented out because of page limits ... I have added back in
> our short version of that section ... but do we want to include it
> or not?
>
> - figure out how to stick 10 authors into the new format ...
>
> - figure out the correct terms, index, keywords ... blah blah blah
>
> Please reply and add things to this list.
>
> Laurie
>
>
>
>
> +-------------------------------------------------------------+
> | Laurie Hendren, Professor, School of Computer Science |
> | McGill University |
> | 318 McConnell Engineering Building tel: (514) 398-7391 |
> | 3480 University Street fax: (514) 398-3883 |
> | Montreal, Quebec H3A 2A7 hendren@cs.mcgill.ca |
> | CANADA http://www.sable.mcgill.ca/~hendren |
> | http://wwww.sable.mcgill.ca http://aspectbench.org |
> +-------------------------------------------------------------+
>
>
Received on Fri Apr 8 15:48:22 2005