itaps-parallel mesh adaptivity example for SC08

Onkar Sahni osahni at scorec.rpi.edu
Tue Nov 11 22:44:27 CST 2008


Lori,

I appreciate your efforts in looking into mesh-adaptivity example and
providing detailed feedback. See my comments below majority of which are
fixes (my comments are inlined and starts with **, it seems like my email
client has jumbled some text from your message while replying to it):

>  - can you move the note/iMeshMA.README.SC08 up one level and rename it
> HandsOn_Session.txt to be consistent with the other exercises

** Fixed, see HowTo.txt and HandsOn_Session.txt under:
/usr/local/itaps/mesh_refinement/

>   - can you move the main.cc file and main Makefile up as well?  In
> addition, I think it would be good if you could list the main function
as the first (and maybe only) thing in the file.  By the time the
students get to this point they will recognize many of the operations
and functions that you use there, which will help them get comfortable
with this example.

** This will be difficult for me as it will require (non-trivial) changes
in Makefile and (iMeshMA's) svn repository structure. I do not plan to
talk about main.cc at all. I will just focus on three iMeshP calls as
mentioned in HandsOn_Session_iMeshMA.txt.

>   - I find the text description in the notes to be a little sparse.  I
> was able to follow the directions and run all the test cases, so that
was good. (As a general comment, we should remind folks how to set the
display for visualization of results.)    You talk about specific iMeshP
functions that you focus on, but there's nothing really to do except go
look at the function (also the first one has a typo - should be
> pushTagsEnt).  Can you tell me what you have in mind?

** I plan to go through the code and show how three specific iMeshP calls
are used. I plan to spent some time in description and discussions related
to these three iMeshP calls.

>  - I like the description of the overall skeleton of the mesh adapt
> service, but think that might be better earlier in this file (before the
description of the particular iMeshP functions).  I'm not sure I see a
need for ~~ vs ~... I'd just list all the steps as equals

** Fixed.

>  - I don't understand the notation for detailed iMeshP function
> listings.. what's the difference between +++, single quotes, and === ...
also, you use a lot of iMesh functions as well, but don't mention
them... why not?   As a minor comment, the caps/lowercase letters aren't
consistent with the header files.

** Fixed (I want to keep focus on iMeshP for various reasons).

>  - It would be useful to explicitly say what you want the students to
> get out of this exercise.  Are there things they should particularly be
looking for in the functions you highlight..  are you hoping they'll get
a sense of how to write/use a mesh adapt service?

** This can cover many things, for example, basic understanding of three
iMeshP calls, a basic demonstration of interoperability, promotion of a
mesh-adapt service from ITAPS to scientific computing app. and showing a
mesh modification app. how iMeshP fulfills it needs for parallel mesh
adaptivity... we can discuss more on this tomorrow during our tele-con.

I am planning to talk/discuss during mesh adaptivity example that will
allow me to describe things. This is the reason why I wanted to get an
idea on time distribution for each hands-on session that will help me to
plan my discussions.

If I missed on any of your concerns/comments then please let me know.

Thanks,
Onkar

>
> Perhaps we can talk about this more tomorrow during the telecon...
>
> Lori
>
> Onkar Sahni wrote:
>> I have updated mesh adaptivity example (based on useful feedback from you
>> guys and updates to database libraries) under:
>> /usr/local/itaps/mesh_refinement
>> (see /usr/local/itaps/mesh_refinement/notes/iMeshMA.README.SC08 to
compile, to run and for other details).
>> Please feel free to try it out and let me know your comments.
>> Thanks,
>> Onkar
>>> I have setup mesh adaptivity example under:
>>> /usr/local/itaps/mesh_refinement
>>> See /usr/local/itaps/mesh_refinement/notes/iMeshMA.README.SC08 to
>> compile
>>> and run (and other details).
>>> Source code (clean version) is under:
>>> /usr/local/itaps/mesh_refinement/src
>>> A build is provided under:
>>> /usr/local/itaps/mesh_refinement/build
>>> Please try it out and let me know your comments.
>>> Thanks,
>>> Onkar
>
>







More information about the itaps-parallel mailing list