[Trilinos-Users] Possible Bug with MPI Communication

[John Reinert]

Hello

I am a graduate student at the University of Minnesota.  I've recently
started using AztecOO.  I've found that for a simple test case, the result
I obtain is affected by the number of processors I run on.  I've attached
the code (test_sine.cpp), which solves a simple 1D advection/diffusion
problem.  I've also included the Makefile to show which libraries I am
compiling with.

I am using Intel 16.0, openmpi 3.0,  and Trilinos 12-12.

After running the code with 1, 2 and 3 processors, I obtain the results
shown in "plot.pdf".  As you can see, the results show small oscillations
where I assume the processor division is located.  Is this a known issue
with Trilinos 12-12?  Should I revert back to an older version?

Best regards,
John
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://trilinos.org/pipermail/trilinos-users/attachments/20180204/5cf6e6c1/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: plot.pdf
Type: application/pdf
Size: 4088 bytes
Desc: not available
URL: <https://trilinos.org/pipermail/trilinos-users/attachments/20180204/5cf6e6c1/attachment.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Plotter.m
Type: text/x-objcsrc
Size: 307 bytes
Desc: not available
URL: <https://trilinos.org/pipermail/trilinos-users/attachments/20180204/5cf6e6c1/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Makefile
Type: application/octet-stream
Size: 1400 bytes
Desc: not available
URL: <https://trilinos.org/pipermail/trilinos-users/attachments/20180204/5cf6e6c1/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test_sine.cpp
Type: text/x-c++src
Size: 4313 bytes
Desc: not available
URL: <https://trilinos.org/pipermail/trilinos-users/attachments/20180204/5cf6e6c1/attachment-0001.bin>