Different Eigenvalues on Different Machines

[Samuel_M_1]

Hello, 

I am computing eigenvalues on different machines. With the exact same binary and the same shared libraries, I obtain different eigenvalues for near zero entries when I run on different machines. Is this expected behavior?

On a 64 core node:

   -2.7756e-17
   4.2301e-46
   3.4694e-18
   1.7984e-02
   9.4446e-01

On a 32 core node:

   1.8526e-47
   3.4694e-18
   2.7756e-17
   1.7984e-02
   9.4446e-01

I suspect that MKL is auto-threading with different numbers of cores and this is causing the floating point operations to execute in different orders, but would like to make sure. 

[Gennady_F_Intel · Accepted solution]

yes, You are right. this is the expected behavior. to eliminate the problem you can try to apply CNR mode.