[Mpi3-tools] Final MPI-3 Tools BOF Slides

Martin Schulz schulzm at llnl.gov
Sat Nov 6 20:58:46 CDT 2010

Hi Bronis,

On Nov 6, 2010, at 6:00 PM, Bronis R. de Supinski wrote:

> Martin:
> Re:
>> Thanks to all for the comments and the feedback. I tried to
>> include all comments (more detailed responses summarized
>> below). Attached are the final slides that I will send to Rich.
> This version is fine but I have no idea what you meant by
> "more detailed responses summarized below" -- perhaps you
> forgot to include something into your email.

This is very strange - somehow Apple Mail or Exchange put
all text in my email that was below the attachment into a 
separate HTML attachment, which should be attached to
the original email (at least that's how my outfolder has it).
I have not seen this before - the text is appended below
again just to make sure (not that it is overly important at
this point).



> I think the slides are too full. Maybe I am missing what the BoF session is for. Am I correct that it is meant to reach out beyond the Forum?

I tried to cut it down even more. I think we need a bit more than
in your proposal, but I understand your direction.

> I think it is best to have the user relate the info to his own experience. That's why in my proposal I just wrote "it like PAPI just for 'Software Counters' and with read--write access". A single sentence and whoever is listening should have a clear view on what we want to accomplish. If we directly want comments on the internals, people have to read the draft first anyway.

I added the PAPI example back.

> Also. I would not cite the direct link to the drafts. People will not be able to memorize the whole url, so I think the critical info is: svn.mpi-forum.org and "Tools WG". Then, I would place a link to each draft prominently on the beginning of the page (maybe below "List of …")

Good point - I changed the slides and I will change the Wiki


> The animated MPIT slide is good. I tweaked the animation
> slightly in the attached so that each examples text box
> appears just after the relevant animation.

That actually changed the meaning - the example had the
setup phase to the left and then an example for a performance
variable on the right. Control variables were not in. I changed
the text to avoid this misunderstanding when Rich is presenting.

> Also, I thought we agreed that we would not use hidden slides
> as it is confusing. In any event, I think you need some words

Those were really intended only for the Web version only.

> on MPIT even with the animation. Further, the hidden MPIR
> texxt slide was almost identical to the unhidden one. I unhid
> all of the slides in the attached and merged the two MPIR
> text slides. A good tweak would be to animate the MPIR process
> figure but that would be above and beyond.

If we don't want any hidden slides - the last one seems too much.
I got rid of it - people can look at the document.

> Overall the current version does put us a couple slides over
> the request. However, it goes over due to the graphic slides,
> whichwill help to make the talk more interesting so I would
> hope that Rich would be OK with that.

We are now 1 slide over, so I hope this is OK.


> I really like the examples and animation you added to slide 2! Now that you have that, I think you can remove some of the text from slide 3 to make it less dense without losing information, e.g. the examples of variables. I edited slide 3 so that it's somewhat less dense but still something that I think Rich can read from and get the information we want across (attached slide 3).

Slide 3 was not supposed to be shown and only intended for the
web version. Since both you and Bronis seemed to not like this,
I added your modified slide as an unhidden one (in which case
I like the reduced text).

> Also, there was a typo in slide 3 that I think you will want to fix in any case: "if the MPI implementation support this" change support to supports.

That's in there now.

Martin Schulz, schulzm at llnl.gov, http://people.llnl.gov/schulzm
CASC @ Lawrence Livermore National Laboratory, Livermore, USA

> Bronis
> _______________________________________________
> Mpi3-tools mailing list
> Mpi3-tools at lists.mpi-forum.org
> http://BLOCKEDlists.mpi-forum.org/mailman/listinfo.cgi/mpi3-tools

Martin Schulz, schulzm at llnl.gov, http://people.llnl.gov/schulzm
CASC @ Lawrence Livermore National Laboratory, Livermore, USA

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpi-forum.org/pipermail/mpiwg-tools/attachments/20101106/f720d1a0/attachment-0001.html>

More information about the mpiwg-tools mailing list