- itaps-parallel - mailman3.cels.anl.gov

Going to commit examples...
by Tim Tautges 12 Nov '08

12 Nov '08

Hi all, I've built the HelloMesh examples in f77, f90, c, and c++. I'll commit them to the rpi repo shortly. Lori, I redid the configure stuff that you had done, to make it a bit easier. You can now enable/disable each language individually, point to your imesh/igeom/etc. install directory, and off it goes. I also put all the languages in the same directory for each test, since I think that's easier to work with. I'll build these on the mesh machine after committing them. - tim -- ================================================================ "You will keep in perfect peace him whose mind is steadfast, because he trusts in you." Isaiah 26:3 Tim Tautges Argonne National Laboratory (tautges(a)mcs.anl.gov) (telecommuting from UW-Madison) phone: (608) 263-8485 1500 Engineering Dr. fax: (608) 263-4499 Madison, WI 53706

1 0

[Fwd: [Fwd: Reminder - phone teleconference tomorrow]]
by Lori A. Diachin 11 Nov '08

11 Nov '08

Hi All, A quick reminder that we have a telecon tomorrow (Wednesday, Nov 12) to discuss the status of the hands-on exercises for the SC08 tutorial along with any issues regarding the iMesh and iMeshP interfaces. It's scheduled for 11 am pacific. This will probably be close to our last chance to chat before the tutorial, so please be prepared with any issues that you feel need to be discussed. The LLNL telecon number is 866-914-3976 and the passcode is 662912# Lori

1 0

iMeshP_getNumOfTypeAll and iMeshP_getNumOfTopoAll
by Jason Kraftcheck 11 Nov '08

11 Nov '08

iMeshP_getNumOfTypeAll and iMeshP_getNumOfTopoAll are too general to implement efficiently. And the behavior is unclear. Consider the following case: - 2 processors - a vertex is on the interface, and therefore has a copy on each - in one proc, the 'entity_set_handle' does not contain the vertex - on the other, it does Is this vertex counted? Either way, it is not possible for every processor to send just a count. It is necessary to gather data from each processor for each entity for which that processor has a copy to determine if/when to count vertices like the above. Further, it isn't even possible to provide an optimized case for the root set (e.g. count owned entities on each proc) because there's no guarantee that because the root set is passed in one processor that it is on any other processor. This seems like yet another case of an ill-considered and overly-complex API. If the intention is to get global counts (e.g. entity_set_handle == root_set), then I think we should just drop the entity_set_handle argument. If there is a realistic case for having the entity_set_handle argument, then a) We should have functions both with and without it so that the without case doesn't incur the horrible performance penalty of the 'with' case (or alternately mandate that if it is the root set on one proc, it must be the root set on all procs.) b) It would be nice if whomever does have a use case for the 'with' case share it, so that we have some basis upon which to decide the 'all vs any' thing. - jason

2 2

Re: itaps-parallel Clarification of iMeshP_getNumPartNborsArr
by Jason Kraftcheck 10 Nov '08

10 Nov '08

Carl Ollivier-Gooch wrote: > Jason Kraftcheck wrote: >> Does iMeshP_getNumPartNborsArr return the sum of the result of >> iMeshP_getNumPartNbors for each part, or the size of the difference of >> the >> union of the results of iMeshP_getPartNbors for each part with the input >> list? For example, if I have three parts arranged such that each >> neighbors >> both other parts and I pass in two of the parts to >> iMeshP_getNumPartNborsArr, is the result: >> 1) 1 because there is one part not in the input list that is adjacent to >> at least one part in the input list >> 2) 4 because each part in the input list has two adjacent parts > > I thought it was > > 3) an array that lists the number of part neighbors for each part (so > (2,2) in this case). But my understanding may be incompatible with the > call list. > That looks to be correct: it returns an array of results. Apologies for any confusion. I was mixing up this function with iMeshP_getPartNborsArr (and assuming that given the similarity of names, that the 'Num' function would be returning the size of the results from other.) However, as iMeshP_getPartNborsArr does not return an offset list, the behavior of that one is still unclear. Are these functions to be used together, or does getPartNborsArr return the union of the results for each part (or the intersection)? - jason

2 1

Re: itaps-parallel Clarification of iMeshP_getNumPartNborsArr
by Carl Ollivier-Gooch 10 Nov '08

10 Nov '08

Jason Kraftcheck wrote: > Does iMeshP_getNumPartNborsArr return the sum of the result of > iMeshP_getNumPartNbors for each part, or the size of the difference of the > union of the results of iMeshP_getPartNbors for each part with the input > list? For example, if I have three parts arranged such that each neighbors > both other parts and I pass in two of the parts to > iMeshP_getNumPartNborsArr, is the result: > 1) 1 because there is one part not in the input list that is adjacent to > at least one part in the input list > 2) 4 because each part in the input list has two adjacent parts I thought it was 3) an array that lists the number of part neighbors for each part (so (2,2) in this case). But my understanding may be incompatible with the call list. Carl -- ------------------------------------------------------------------------ Dr. Carl Ollivier-Gooch, P.Eng. Voice: +1-604-822-1854 Associate Professor Fax: +1-604-822-2403 Department of Mechanical Engineering email: cfog(a)mech.ubc.ca University of British Columbia http://www.mech.ubc.ca/~cfog Vancouver, BC V6T 1Z4 http://tetra.mech.ubc.ca/ANSLab/ ------------------------------------------------------------------------

1 0

Clarification of iMeshP_getNumPartNborsArr
by Jason Kraftcheck 10 Nov '08

10 Nov '08

Does iMeshP_getNumPartNborsArr return the sum of the result of iMeshP_getNumPartNbors for each part, or the size of the difference of the union of the results of iMeshP_getPartNbors for each part with the input list? For example, if I have three parts arranged such that each neighbors both other parts and I pass in two of the parts to iMeshP_getNumPartNborsArr, is the result: 1) 1 because there is one part not in the input list that is adjacent to at least one part in the input list 2) 4 because each part in the input list has two adjacent parts - jason

1 0

Re: itaps-parallel FronTier tutorial
by Mark Miller 10 Nov '08

10 Nov '08

Hi All, FYI, I tested a modification to Ryan's Makeile to use '-static' on both creation of .o (compile) and creation of executable (link). That resulted in odd error for '-lgcc_s' (which I assume is probably needed for shared libs maybe?). After I removed '-lgcc_s' and added '- lz' (which is needed because hdf5 lib is compiled with zlib depdnence), I got a static executable but with following warning... /usr/local/hdf5serial/lib/libhdf5.a(H5FDstream.o): In function `H5FD_stream_open': H5FDstream.c:(.text+0x125e): warning: Using 'gethostbyname' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking However, the static executable behaved identically to the shared executable. I still get mcmiller@mesh:~$ ./HelloFronTier error type 3 in: iMesh_load Mark On Sun, 2008-11-09 at 19:44 -0800, Mark Miller wrote: > Hi Ryan and Martin, > > Ok, I tried it and got further after finding hdf5 libs... > > ./HelloFronTier: error while loading shared libraries: libiMesh.so.0: cannot open shared object file: No such file or directory > mcmiller@mesh:~$ env LD_LIBRARY_PATH=/usr/local/itaps/MOABser/lib ./HelloFronTier > ./HelloFronTier: error while loading shared libraries: libhdf5.so.0: cannot open shared object file: No such file or directory > mcmiller@mesh:~$ env LD_LIBRARY_PATH=/usr/local/itaps/MOABser/lib:/usr/local/hdf5serial/lib ./HelloFronTier > error type 3 in: iMesh_load > > mcmiller@mesh:~$ env LD_LIBRARY_PATH=/usr/local/itaps/MOABser/lib:/usr/local/hdf5serial/lib ./HelloFronTier /usr/local/itaps/FronTier_tutorial/vtk_files/circle.vtk -o vtk > Usage: example -o output > > Also, as I am sure you are probably already thinking, it is probably > best to force static linking whenever possible at least where tutorials > are concerned as you can eliminate issues with finding shared libs. > > Mark > > > On Fri, 2008-11-07 at 20:21, isenburg1(a)llnl.gov wrote: > > hi ryan, > > > > i tried running your tutorial. there is some issue with libhdf5.so.0 > > libraries. are they missing or is an obvious path missing? > > > > isenburg@mesh:~/FronTier_tutorial$ ./HelloFronTier > > ./HelloFronTier: error while loading shared libraries: libiMesh.so.0: > > cannot open shared object file: No such file or directory > > > > isenburg@mesh:~/FronTier_tutorial$ echo $LD_LIBRARY_PATH > > /usr/local/itaps/FMDB_iMeshP/lib/ > > > > isenburg@mesh:~/FronTier_tutorial$ export > > LD_LIBRARY_PATH=/usr/local/itaps/MOABser/lib:${LD_LIBRARY_PATH} > > > > isenburg@mesh:~/FronTier_tutorial$ echo $LD_LIBRARY_PATH > > /usr/local/itaps/MOABser/lib:/usr/local/itaps/FMDB_iMeshP/lib/ > > > > isenburg@mesh:~/FronTier_tutorial$ ./HelloFronTier > > ./HelloFronTier: error while loading shared libraries: libhdf5.so.0: > > cannot open shared object file: No such file or directory > > > > cheers > > > > martin -- Mark C. Miller, Lawrence Livermore National Laboratory email: mailto:[email protected] (M/T/W) (925)-423-5901 (!!LLNL BUSINESS ONLY!!) (Th/F) (530)-753-8511 (!!LLNL BUSINESS ONLY!!)

1 0

FronTier 2D tutorial in /usr/local/itaps/FronTier_tutorial
by Ryan Kaufman 10 Nov '08

10 Nov '08

I have moved the FronTier tutorial code to /usr/local/itaps/FronTier_tutorial. Perhaps someone can try it out? I did something neat using iterators in a for loop. I copied part of the README from martin. Hope that's OK. I might make small changes or tweaks this weekend, but I think it's ready. I will try to post a 3D example as well, but that depends on 3d actually working. -- ~Ryan Kaufman

3 3

testing the Smooth tutorial
by Martin Isenburg 10 Nov '08

10 Nov '08

hello, does anyone have a few minutes to spare to try out the "Smooth" tutorial on /usr/local/itaps/Smooth? let me know if you have any comments ... martin

2 1

FronTier tutorial
by Ryan Kaufman 09 Nov '08

09 Nov '08

I'm finishing up the FronTier 2D tutorial example, which I will install tonight to /usr/local after I tidy up the works a bit. So, the user has to set a tag on each point with the velocity at each time step. He can do this by calling getEntities, and then looping over the resulting array; or he can use an entity iterator. Do we have a preference on this? I think it may be neat to use the iterator. -- ~Ryan Kaufman

3 3