Different stability results in serial and parallel running
Xiao Chengnian retported this
I have been running nektar++-5.0.2 and trying the 3D stability analysis for the channel flow
as described in the specific nektar tutorial. I used the session file "PPF_R10000_3D.xml"
provided under the /Tests subdirectory of IncNavierStokesSolver, and obtained the same
complex eigenvalues 0.518448+/-26.6405i as the tutorial when I ran the problem on
a single processor. However, when I executed the same session on multiple (4) processors,
the two eigenvalues upon convergence were different, namely -6.957 +/-1.34189i.I can repeat this problem in the current master.
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information